[{"data":1,"prerenderedAt":1337},["ShallowReactive",2],{"categories-init":3,"stack-local-llm-dev-stack":4,"stack-res-local-llm-dev-stack":1336},true,{"stack_id":5,"slug":6,"name":7,"tagline":8,"long_description":9,"key_features":10,"use_cases":18,"pros":24,"cons":29,"cover_image_url":35,"scores":36,"options":50,"additions":564,"addition_groups":1029,"option_groups":1054,"multi_select_option_types":1055,"tools_by_category":1057,"related_stacks":1121,"faqs":1280,"pricing":1299,"system_requirements":35,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"published_date":172,"last_updated_date":1331,"seo_meta":1332},241,"local-llm-dev-stack","Local LLM Dev Stack","A coding agent on a model that runs on your own machine, with no code sent to a provider.","This stack runs a coding agent against **a model on your own hardware**. A local runtime serves an open-weight model on the machine, and the agent talks to that address in place of a provider's API, so prompts, code, and file contents never leave it. There is no subscription and no per-token bill, it works offline, and nothing changes when a vendor changes its prices or its rules. The price is quality and speed: an open model that fits one graphics card is weaker on long, multi-file tasks than the hosted models these agents were designed around.\n\nThe runtime is **the piece you operate**. Ollama is the default: a background service with a model library, and a launch command that starts OpenCode, Claude Code, or Codex already pointed at a local model. LM Studio does the same job from a desktop app and uses Apple's MLX engine on Apple Silicon Macs, llama.cpp is the bare engine for people who want control over every setting, and Unsloth also fine-tunes the models it runs. All four expose the same kind of local server, so the agent does not care which one is behind it.\n\nOpenCode is the default agent because it is open source, tied to no model vendor, and documents these runtimes itself. Cline, Qwen Code, Codex, and Claude Code work as well, each through a setting or an endpoint that its vendor or the runtime documents, and several agents can share one local model. **The model matters more than the agent.** Qwen's coding models are the default, with Mistral's Devstral, Google Gemma, and DeepSeek as alternatives, and all of them need a large context window to work with an agent, far above what the runtimes set out of the box.\n\nBy default everything runs on **your own computer**, which has to stay on while the agent works. A common layout splits it in two: the model on a desktop with a strong graphics card, and the agent on a laptop that reaches it over the home network, or over Tailscale when you are away. OpenCode's web interface serves the same sessions to a phone browser. There is no vendor cloud to hand work to, which is the point of the stack. The model can also live on a rented GPU server, which keeps it available around the clock at a monthly price well above a coding subscription.\n\nCode lives on GitHub by default, or on GitLab when source control has to stay on your own infrastructure too. A model gateway such as OpenRouter can be added as a fallback for the tasks a local model can't finish, and a faster terminal, an AI code reviewer, and CI\u002FCD are common additions.",[11,12,13,14,15,16,17],"A coding agent working against a model on your own hardware, with no code sent to a provider","A choice of local runtime: Ollama by default, or LM Studio, llama.cpp, or Unsloth","One agent or several on the same model: OpenCode by default, with Cline, Qwen Code, Codex, or Claude Code","Open coding models that fit one graphics card: Qwen's coder models, Devstral, Gemma, and DeepSeek","Works offline, with no subscription and no per-token cost","The model can sit on a second, stronger machine and be reached over your network","Hosted models through a gateway as an optional fallback for hard tasks",[19,20,21,22,23],"Working on code that is not allowed to leave your machine or your company's network","Coding on a plane, a train, or anywhere without a reliable connection","Routine edits, tests, and refactors without spending tokens","Comparing open coding models side by side on your own repository","One shared model machine that several developers' agents connect to",[25,26,27,28],"Nothing leaves your hardware: prompts, code, and file contents stay local","No subscription, no usage limits, and no per-token bill","Every layer is replaceable: the runtime, the model, the agent, and the git host","The same local model also serves chat interfaces and other tools on your network",[30,31,32,33,34],"Open models on one graphics card are weaker than hosted frontier models on long, multi-file tasks","Real hardware is needed: about 24 GB of graphics memory, or a Mac with 32 GB or more, for the coding models","The context window has to be raised well above the runtime's default, which costs memory and speed","No vendor cloud sessions or phone app: away from the desk, a machine you own has to stay on","Claude Code and Codex reach local models through compatibility layers, so some of their features are missing",null,{"popularity":37,"learning_curve":40,"flexibility":42,"performance":45,"portability":48},{"score":38,"reasoning":39},4,"Ollama and OpenCode are both widely recognised names, and running a coding agent against a locally hosted Qwen model is a common pairing in local-model communities. It is better known among AI-tooling developers than outside them.",{"score":38,"reasoning":41},"Beyond learning the agent, there is a runtime to operate, a model to size against the hardware, and a context window to raise before anything works. Ollama's launch command shortens the first run, but getting good results takes tuning.",{"score":43,"reasoning":44},5,"Four runtimes, five agents, any open-weight model, and two git hosts, each replaceable without touching the rest, with a hosted model available through a gateway when needed.",{"score":46,"reasoning":47},2,"A model that fits one graphics card answers more slowly than a hosted one and handles long, multi-file tasks less reliably. Contained tasks go well, and results depend heavily on the hardware and the context size.",{"score":43,"reasoning":49},"Every part is open source or an open-weight file, the runtimes speak the same local API, and nothing is tied to an account or a plan. The setup moves to another machine by copying models and settings.",{"model_inference":51,"coding_agent":224,"llm":373,"version_control":500},{"name":52,"tools":53,"descriptions":215,"aliases":220,"see_all":221},"Model Inference",[54,109,144,173],{"tool_id":55,"name":56,"slug":57,"tooltip_description":58,"logo_url":59,"logo_bg":60,"pricing_model":61,"learning_curve_score":65,"popularity_score":43,"hosting_assignment_type":66,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":68,"subcategory":72,"categories":76,"subcategories":79,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":82,"aliases":100,"score_reasonings":101,"published_date":107,"last_updated_date":108},210,"Ollama","ollama","The most widely used way to run open-weight LLMs locally — one command downloads and serves models like Llama, Qwen, DeepSeek, GLM, and MiniMax through an OpenAI-compatible API, with an optional paid Ollama Cloud tier for larger models than local hardware can handle.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Follama.svg","white",{"slug":62,"display_name":63,"description":64},"freemium","Freemium","A free tier is available; additional features, usage limits, or managed hosting require a paid plan.",1,"deployable","open",{"category_id":69,"name":70,"slug":71},21,"AI Infrastructure","ai-infrastructure",{"subcategory_id":73,"name":74,"slug":75},60,"AI Runtime & Serving","ai-runtime-serving",[77],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},0,[80],{"subcategory_id":73,"name":74,"slug":75,"category_id":69,"is_primary":3,"display_order":78},false,[83,88,92,96],{"tag_id":84,"name":85,"slug":86,"tag_type":87},11,"Open Source","open-source","feature",{"tag_id":89,"name":90,"slug":91,"tag_type":87},12,"Self-hostable","self-hostable",{"tag_id":93,"name":94,"slug":95,"tag_type":87},13,"Free Tier","free-tier",{"tag_id":97,"name":98,"slug":99,"tag_type":87},16,"AI-powered","ai-powered",[],{"learning_curve":102,"flexibility":103,"performance":104,"popularity":105,"portability":106},"A single install command and a single command to pull and run a model make it the lowest-friction way to try a local LLM, with no configuration required to get started.","100+ supported models, an OpenAI-compatible API for drop-in tooling reuse, and both local and cloud execution modes give it very broad applicability across workflows.","Automatic hardware tuning and support for the latest open-weight models keep it competitive for local inference, though it is not purpose-built for high-throughput production serving the way dedicated inference servers are.","The most widely used local LLM runtime by a clear margin, with 178K+ GitHub stars, 52 million monthly downloads, and 2.5 billion+ cumulative downloads.","MIT-licensed, runs on macOS, Windows, and Linux across Apple Silicon, NVIDIA, and AMD hardware, and models are entirely self-hosted with no forced cloud dependency.","2026-09-27","2026-09-28",{"tool_id":110,"name":111,"slug":112,"tooltip_description":113,"logo_url":114,"logo_bg":115,"pricing_model":116,"learning_curve_score":65,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":117,"subcategory":118,"categories":119,"subcategories":121,"flexibility_score":38,"performance_score":38,"portability_score":38,"is_featured":81,"tags":123,"aliases":136,"score_reasonings":137,"published_date":143,"last_updated_date":143},299,"LM Studio","lm-studio","Desktop app for downloading and running open-weight language models locally on macOS, Windows, and Linux, with a chat interface and a local OpenAI-compatible API server. It is free for personal and work use.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Flm-studio.png","dark",{"slug":62,"display_name":63,"description":64},{"category_id":69,"name":70,"slug":71},{"subcategory_id":73,"name":74,"slug":75},[120],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},[122],{"subcategory_id":73,"name":74,"slug":75,"category_id":69,"is_primary":3,"display_order":78},[124,125,126,131],{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":127,"name":128,"slug":129,"tag_type":130},25,"Machine Learning","machine-learning","use_case",{"tag_id":132,"name":133,"slug":134,"tag_type":135},43,"Cross-platform","cross-platform","platform",[],{"learning_curve":138,"flexibility":139,"performance":140,"popularity":141,"portability":142},"A graphical installer, in-app model search, and automatic quantization choices mean a first local model runs within minutes without touching a terminal or config file.","Serves any GGUF or MLX model through OpenAI- and Anthropic-compatible APIs, SDKs, a CLI, and a headless daemon, with MCP tools and remote access through LM Link, though engine internals are not open for modification.","MLX on Apple Silicon and llama.cpp elsewhere give fast single-user inference for the hardware, while throughput for many simultaneous requests is well below datacenter engines such as vLLM.","One of the two default answers, alongside Ollama, whenever someone asks how to run an LLM locally, with a large user base on Mac and Windows beyond professional developers.","Runs on all three desktop operating systems, uses open model formats from Hugging Face, and exposes standard APIs, so models and client code move freely to Ollama or vLLM.","2026-10-01",{"tool_id":145,"name":146,"slug":147,"tooltip_description":148,"logo_url":149,"logo_bg":60,"pricing_model":150,"learning_curve_score":153,"popularity_score":38,"hosting_assignment_type":66,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":154,"subcategory":155,"categories":156,"subcategories":158,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":160,"aliases":165,"score_reasonings":166,"published_date":172,"last_updated_date":35},303,"llama.cpp","llama-cpp","Open-source C\u002FC++ engine that runs GGUF language and vision models on CPUs, Apple Silicon, and consumer GPUs. Its llama-server binary serves them through an OpenAI-compatible API with a built-in web UI.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fllama-cpp.svg",{"slug":151,"display_name":85,"description":152},"open_source","Source code is publicly available and free to use, modify, and distribute. No paid plans from the project itself.",3,{"category_id":69,"name":70,"slug":71},{"subcategory_id":73,"name":74,"slug":75},[157],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},[159],{"subcategory_id":73,"name":74,"slug":75,"category_id":69,"is_primary":3,"display_order":78},[161,162,163,164],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":127,"name":128,"slug":129,"tag_type":130},{"tag_id":132,"name":133,"slug":134,"tag_type":135},[],{"learning_curve":167,"flexibility":168,"performance":169,"popularity":170,"portability":171},"A one-line install and `llama serve -hf` get a model running in minutes, but getting the most out of it means picking a quantization, learning offload and context flags, and sometimes compiling for a specific GPU backend.","Exposes nearly every inference parameter, runs as a CLI, a server, or an embedded library, serves several models through router mode, and speaks both OpenAI and Anthropic API formats.","Best-in-class speed for single-user inference on consumer hardware and Apple Silicon, though batched multi-user throughput on GPU servers trails engines built for it, such as vLLM.","Around 130K GitHub stars and the engine behind much of the local AI ecosystem, well known to anyone running models locally, while developers who only call hosted APIs rarely use it directly.","MIT-licensed with no dependencies, runs on macOS, Windows, Linux, and Android across every major GPU vendor and plain CPUs, and GGUF files work in any compatible runtime.","2026-10-02",{"tool_id":174,"name":175,"slug":176,"tooltip_description":177,"logo_url":178,"logo_bg":115,"pricing_model":179,"learning_curve_score":153,"popularity_score":38,"hosting_assignment_type":180,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":181,"subcategory":185,"categories":189,"subcategories":192,"flexibility_score":38,"performance_score":43,"portability_score":43,"is_featured":81,"tags":195,"aliases":208,"score_reasonings":209,"published_date":107,"last_updated_date":143},236,"Unsloth","unsloth","Open-source desktop app, web UI, and Python library for running and fine-tuning open-weight models locally. It serves models through an OpenAI-compatible API and trains them about twice as fast with roughly 70% less memory than standard pipelines.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Funsloth.png",{"slug":62,"display_name":63,"description":64},"library",{"category_id":182,"name":183,"slug":184},14,"Data & ML Libraries","data-ml-libraries",{"subcategory_id":186,"name":187,"slug":188},67,"Fine-Tuning","fine-tuning",[190,191],{"category_id":182,"name":183,"slug":184,"is_primary":3,"display_order":78},{"category_id":69,"name":70,"slug":71,"is_primary":81,"display_order":65},[193,194],{"subcategory_id":186,"name":187,"slug":188,"category_id":182,"is_primary":3,"display_order":78},{"subcategory_id":73,"name":74,"slug":75,"category_id":69,"is_primary":81,"display_order":65},[196,200,201,202,203,204],{"tag_id":65,"name":197,"slug":198,"tag_type":199},"Python","python","technology",{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},{"tag_id":127,"name":128,"slug":129,"tag_type":130},{"tag_id":205,"name":206,"slug":207,"tag_type":130},39,"Data Science","data-science",[],{"learning_curve":210,"flexibility":211,"performance":212,"popularity":213,"portability":214},"The prebuilt Colab and Kaggle notebooks make a first fine-tune approachable, but producing a genuinely good model still requires dataset preparation, hyperparameter tuning, and evaluation skill.","Covers LoRA, QLoRA, full fine-tuning, pretraining, and RL methods across many model families, though it is purpose-built for fine-tuning rather than a general training framework.","Hand-written fused kernels deliver roughly 70% memory reduction and about 2x faster training with no accuracy loss, best-in-class for efficient fine-tuning.","One of the most-starred projects in the fine-tuning space and a common default in tutorials, with widely used quantized model releases on Hugging Face.","Apache 2.0 licensed, self-hostable on any compatible GPU, and exports adapters to GGUF, Ollama, and vLLM, so trained models run anywhere with no lock-in.",{"ollama":216,"lm-studio":217,"llama-cpp":218,"unsloth":219},"Ollama is the default here: a background service that downloads models by name and serves them on a local address. Its launch command starts OpenCode, Claude Code, or Codex already configured for a local model, and Cline has it as a built-in provider. Its default context window is small, and its docs recommend 64,000 tokens for coding tools, so that setting has to be raised. Its cloud models use the same setup when local hardware runs short. Free and MIT licensed.","LM Studio is the desktop-app route: a model browser that picks a build sized for your memory, a chat window for testing, and a local server with OpenAI-compatible and Anthropic-compatible endpoints. It documents Claude Code and Codex setups itself, and on Apple Silicon it runs models through Apple's MLX engine. LM Link lets a laptop use models loaded on another machine. Free for personal and work use.","llama.cpp is the engine most local tools are built on, used directly. Its server offers OpenAI-compatible and Anthropic-compatible endpoints, downloads models straight from Hugging Face, and exposes every setting: the quantization, how much of the model sits on the graphics card, the context size. OpenCode documents it as a provider. Pick it for full control and a small footprint, at the cost of managing models yourself. Free, MIT licensed.","Unsloth runs local models from a desktop app or a web interface and serves them on an OpenAI-compatible address, with one command that connects OpenCode, Claude Code, or Codex to the loaded model. What sets it apart is training: the same tool fine-tunes a model on your own code or data and then serves the result. Pick it when a tuned model is where this is heading. Free and open source.",{},{"kind":222,"slug":75,"name":74,"href":223},"subcategory","\u002Ftools\u002Fcategories\u002Fai-infrastructure\u002Fai-runtime-serving",{"name":225,"tools":226,"descriptions":364,"aliases":370,"see_all":371},"Coding Agent",[227,258,282,307,338],{"tool_id":228,"name":229,"slug":230,"tooltip_description":231,"logo_url":232,"logo_bg":60,"pricing_model":233,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":234,"subcategory":238,"categories":242,"subcategories":244,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":246,"aliases":250,"score_reasonings":251,"published_date":107,"last_updated_date":257},192,"OpenCode","opencode","Open-source, model-agnostic AI coding agent for the terminal, desktop, and IDE, built by Anomaly. Connects to 75+ providers or your existing Copilot or ChatGPT subscription, with no lock-in to one model vendor.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fopencode.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},8,"Development Tools","development-tools",{"subcategory_id":239,"name":240,"slug":241},49,"AI Coding Agents","ai-coding-agents",[243],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[245],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[247,248,249],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":252,"flexibility":253,"performance":254,"popularity":255,"portability":256},"Getting a first session running is a quick `npm install` plus an API key, but making full use of Build\u002FPlan modes, MCP\u002FLSP integration, and choosing among 75+ providers involves more configuration surface than a single-model vendor tool.","Bring-your-own-model architecture, custom tools and plugins, configurable permissions\u002Fkeybinds\u002Fformatters, and MCP\u002FLSP support give it wide-open extensibility with essentially no imposed workflow.","Output quality tracks whichever underlying model is configured rather than a fixed baseline — with a strong model (e.g. Claude or GPT) selected, it performs competitively with leading vendor agents, but there's no single benchmark number the way there is for a single-model product.","Roughly 198K GitHub stars and 900+ contributors as of mid-2026, reported 7.5M+ monthly users — the most-starred project in the open-source AI coding agent space, ahead of every other independent (non-vendor) entrant.","MIT-licensed, installable via npm\u002FHomebrew\u002FDocker, no code stored on OpenCode's servers, and not tied to any single model provider — among the least locked-in options in the subcategory.","2026-10-12",{"tool_id":259,"name":260,"slug":261,"tooltip_description":262,"logo_url":263,"logo_bg":60,"pricing_model":264,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":265,"subcategory":266,"categories":267,"subcategories":269,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":271,"aliases":275,"score_reasonings":276,"published_date":107,"last_updated_date":257},195,"Cline","cline","Open-source, bring-your-own-key AI coding agent: it started as a VS Code extension and now also runs as a CLI, desktop app, SDK, and browser-based cloud agent, with a distinctive Plan\u002FAct workflow and step-by-step human approval.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fcline.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":239,"name":240,"slug":241},[268],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[270],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[272,273,274],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":277,"flexibility":278,"performance":279,"popularity":280,"portability":281},"Installs as a familiar VS Code sidebar extension and the Plan\u002FAct toggle maps to an intuitive review-then-execute mental model; getting productive requires only an API key.","Bring-your-own-key across 10+ providers, `.clinerules` customization, MCP support, and an SDK for embedding agent behavior into other workflows give it very open-ended extensibility.","Output quality tracks the underlying model selected rather than a fixed baseline, and the default step-by-step approval workflow favors accuracy and reviewability over raw autonomous throughput compared to fully autonomous cloud agents.","66K+ GitHub stars and 8M+ installs make it one of the largest independent (non single-vendor) coding agents, though still behind OpenCode's scale and well behind GitHub Copilot's installed base.","Apache-2.0 licensed, model-agnostic, and available across VS Code, JetBrains, Cursor, Windsurf, and a standalone CLI — no single-editor or single-provider lock-in.",{"tool_id":283,"name":284,"slug":285,"tooltip_description":286,"logo_url":287,"logo_bg":115,"pricing_model":288,"learning_curve_score":153,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":289,"subcategory":290,"categories":291,"subcategories":293,"flexibility_score":38,"performance_score":38,"portability_score":38,"is_featured":81,"tags":295,"aliases":299,"score_reasonings":300,"published_date":107,"last_updated_date":306},198,"Qwen Code","qwen-code","Alibaba's open-source, Apache-2.0 coding agent for the terminal, built alongside the Qwen models but able to use any OpenAI-, Anthropic-, or Gemini-compatible provider or a local model. It also runs as a desktop app and in VS Code, Zed, and JetBrains.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fqwen-code.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":239,"name":240,"slug":241},[292],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[294],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[296,297,298],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":301,"flexibility":302,"performance":303,"popularity":304,"portability":305},"Terminal-first setup is straightforward for anyone used to a CLI agent, and defaulting to Qwen3-Coder removes an initial model-choice decision, though picking a different provider or configuring MCP servers takes more familiarity.","Accepts any OpenAI-compatible endpoint despite its Qwen-first branding, plus MCP integration and multi-IDE support, giving it real configurability beyond its default model pairing.","Qwen3-Coder's 69.6% SWE-bench Verified score is genuinely competitive, ahead of the Gemini build the tool was forked from and within range of Claude Sonnet-tier results.","27K+ GitHub stars places it among the more established vendor-CLI coding agents, though well behind OpenCode, Cline, and GitHub Copilot in overall community size.","Apache-2.0 licensed and works with any OpenAI-compatible provider, not just Qwen models, so it avoids hard lock-in even though its documentation and defaults are Qwen-centric.","2026-09-30",{"tool_id":308,"name":309,"slug":310,"tooltip_description":311,"logo_url":312,"logo_bg":115,"pricing_model":313,"learning_curve_score":153,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":314,"subcategory":315,"categories":316,"subcategories":318,"flexibility_score":153,"performance_score":43,"portability_score":46,"is_featured":81,"tags":320,"aliases":326,"score_reasonings":331,"published_date":337,"last_updated_date":306},150,"Codex","codex","OpenAI's AI coding agent that builds features, fixes bugs, reviews code, and runs tests, either in cloud sandboxes or locally, through ChatGPT, an open-source CLI, a desktop app, and IDE extensions.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fcodex.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":239,"name":240,"slug":241},[317],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[319],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[321,322],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},40,"Web","web",[327,329],{"alias":328,"is_primary":81},"ChatGPT",{"alias":330,"is_primary":81},"Astra",{"learning_curve":332,"flexibility":333,"performance":334,"popularity":335,"portability":336},"Basic task creation via natural language is straightforward, but designing effective multi-step agent workflows, debugging sandbox failures, and managing parallel task coordination requires real practice. Score 3.","Cloud-first architecture limits self-hosting; task environments can be customised via setup scripts and env vars, but core agent behaviour is proprietary and not deeply configurable. Score 3.","Powered by codex-1 (o3-based), which leads SWE-bench and real-world coding benchmarks. Parallel execution eliminates sequential bottlenecks. Best-in-class for agentic coding tasks as of 2026. Score 5.","67K GitHub stars on the open-source Codex CLI within its first year; reported 2M+ weekly active users across ChatGPT plans. Strong enterprise adoption but newer than incumbents like Cursor. Score 4.","Task execution is entirely cloud-bound on OpenAI infrastructure. Generated code and PR diffs are portable, but the agent itself cannot be self-hosted or migrated to another provider. Score 2.","2026-05-29",{"tool_id":97,"name":339,"slug":340,"tooltip_description":341,"logo_url":342,"logo_bg":115,"pricing_model":343,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":347,"subcategory":348,"categories":349,"subcategories":351,"flexibility_score":43,"performance_score":38,"portability_score":38,"is_featured":81,"tags":353,"aliases":355,"score_reasonings":358,"published_date":337,"last_updated_date":306},"Claude Code","claude-code","Claude Code is Anthropic's agentic coding tool that runs in your terminal and IDEs, reads your entire codebase, edits files, runs commands, and handles git workflows through natural language. It supports multi-agent orchestration, MCP integrations, and custom automation via CLAUDE.md and hooks.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fclaude.svg",{"slug":344,"display_name":345,"description":346},"paid","Paid","No meaningful free tier — a subscription or one-time purchase is required to use the tool.",{"category_id":235,"name":236,"slug":237},{"subcategory_id":239,"name":240,"slug":241},[350],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[352],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[354],{"tag_id":97,"name":98,"slug":99,"tag_type":87},[356],{"alias":357,"is_primary":81},"Anthropic",{"learning_curve":359,"performance":360,"flexibility":361,"portability":362,"popularity":363},"Agentic CLI workflow is new for most developers; hooks, context, and permissions take time.","Lightweight terminal CLI; very responsive with no significant overhead beyond API latency.","Hooks, custom slash commands, MCP servers, and the Agent SDK provide deep extensibility.","CLI tool with standard terminal workflow; hooks and commands are reasonably portable.","The most widely discussed AI coding agent as of 2025\u002F2026; very high developer mindshare.",{"opencode":365,"cline":366,"qwen-code":367,"codex":368,"claude-code":369},"OpenCode is the default here: an open-source agent with no model vendor behind it, in the terminal, a desktop app, and a web interface. Its docs cover Ollama, LM Studio, and llama.cpp as providers, and Ollama's launch command sets it up in one step. Ollama's docs put its context need at 64,000 tokens or more, and its plan and build agents can each use a different local model. Free, MIT licensed.","Cline is an agent that lives in VS Code and other editors, with Ollama and LM Studio as built-in providers. Its docs are frank about local use: a compact-prompt setting for small models, and memory tiers from 16 GB for small quantized models to 64 GB and more for large ones. A good pick when you want to watch and approve each edit in the editor. Free and open source, and with a local model there is nothing else to pay for.","Qwen Code is Alibaba's open-source terminal agent, built around the Qwen models that are this stack's default weights. It reaches a local runtime as an OpenAI-compatible provider with a local address, and has a timeout setting for slow local servers. The natural pairing when the model is a Qwen coder, since agent and model come from the same team. Free and open source.","OpenAI's Codex CLI has a mode for open models that points it at a local Ollama or LM Studio server, and both runtimes document the setup. It keeps Codex's approval flow while the model runs on your machine. Codex is built around OpenAI's own models, so a local model gets a prompt tuned for a different family, and it is worth testing on a real task before settling on it. The CLI is free and open source.","Claude Code can run against a local model through the Anthropic-compatible endpoints that Ollama, LM Studio, and llama.cpp provide, and Ollama's launch command or Unsloth's start command sets it up. Anthropic supports Claude Code with Claude models only, so this route belongs to the runtimes: Ollama's docs note that hosted web search and some tool controls are not fully supported. Pick it when you already work in Claude Code and want the same interface offline.",{},{"kind":222,"slug":241,"name":240,"href":372},"\u002Ftools\u002Fcategories\u002Fdevelopment-tools\u002Fai-coding-agents",{"name":374,"tools":375,"descriptions":491,"aliases":496,"see_all":497},"LLM",[376,409,443,467],{"tool_id":377,"name":378,"slug":379,"tooltip_description":380,"logo_url":381,"logo_bg":115,"pricing_model":382,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":383,"subcategory":386,"categories":390,"subcategories":392,"flexibility_score":43,"performance_score":43,"portability_score":43,"is_featured":81,"tags":394,"aliases":398,"score_reasonings":403,"published_date":337,"last_updated_date":107},162,"Qwen","qwen","Alibaba's LLM family, from sub-1B models to a 2.4-trillion-parameter flagship, covering reasoning, coding, vision, and long context, mostly under Apache 2.0 open weights with a hosted API on Alibaba Cloud.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fqwen.svg",{"slug":62,"display_name":63,"description":64},{"category_id":384,"name":374,"slug":385},19,"llm",{"subcategory_id":387,"name":388,"slug":389},52,"Open-weight","open-weight",[391],{"category_id":384,"name":374,"slug":385,"is_primary":3,"display_order":78},[393],{"subcategory_id":387,"name":388,"slug":389,"category_id":384,"is_primary":3,"display_order":78},[395,396,397],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},{"tag_id":89,"name":90,"slug":91,"tag_type":87},[399,401],{"alias":400,"is_primary":3},"Alibaba",{"alias":402,"is_primary":3},"QwQ",{"learning_curve":404,"flexibility":405,"performance":406,"popularity":407,"portability":408},"Open weights + OpenAI-compatible API means onboarding is straightforward for any developer familiar with the OpenAI SDK. Ollama makes local inference a one-command setup. Navigating the large model family and choosing the right variant adds some complexity, but the barrier is low overall.","Apache 2.0 licence, sizes from 0.6B to 480B+, dense and MoE architectures, specialised coding and vision variants, toggleable reasoning mode, and multiple access paths (self-hosted, DashScope, third-party APIs). Maximum flexibility for any use case.","Qwen3-Coder ranks among the very best open-weight coding models; Qwen3 MoE variants compete with GPT-4-class models on benchmarks; QwQ-32B rivals leading reasoning specialists; Qwen-VL leads vision-language open-weight benchmarks. Across all domains, Qwen is at or near the open-weight frontier.","Qwen3 repo has ~27K GitHub stars; Qwen3-Coder ~16K; QwenLM org collectively has tens of thousands of stars across repositories. Dominant in Chinese AI ecosystem and rapidly growing global adoption. Strong presence on Hugging Face, Ollama library, and all major inference providers.","Apache 2.0 weights on Hugging Face run on Ollama, vLLM, llama.cpp, LM Studio, and MLX with no licence restrictions. Same models available via DashScope and third-party APIs. Maximum portability — switch providers or go self-hosted any time.",{"tool_id":410,"name":411,"slug":412,"tooltip_description":413,"logo_url":414,"logo_bg":60,"pricing_model":415,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":416,"subcategory":417,"categories":418,"subcategories":420,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":3,"tags":422,"aliases":426,"score_reasonings":437,"published_date":337,"last_updated_date":107},163,"Mistral","mistral","Europe's leading open-weight LLM family, from small edge models to a frontier MoE flagship, available through Mistral's API, its Vibe assistant, the major clouds, or self-hosted.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fmistral.svg",{"slug":62,"display_name":63,"description":64},{"category_id":384,"name":374,"slug":385},{"subcategory_id":387,"name":388,"slug":389},[419],{"category_id":384,"name":374,"slug":385,"is_primary":3,"display_order":78},[421],{"subcategory_id":387,"name":388,"slug":389,"category_id":384,"is_primary":3,"display_order":78},[423,424,425],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[427,429,431,433,435],{"alias":428,"is_primary":3},"Mixtral",{"alias":430,"is_primary":3},"Codestral",{"alias":432,"is_primary":3},"Devstral",{"alias":434,"is_primary":3},"Magistral",{"alias":436,"is_primary":3},"Pixtral",{"learning_curve":438,"flexibility":439,"performance":440,"popularity":441,"portability":442},"The la Plateforme API is OpenAI-compatible, so any developer already familiar with the OpenAI Python SDK can switch with a one-line URL change. Running open-weight models locally via Ollama is also beginner-friendly. The main complexity is the breadth of model choices and understanding trade-offs between MoE and dense architectures.","Unmatched spectrum — from 3B edge models to 675B MoE frontier models; open-weight for full fine-tuning and self-hosting; API for managed access; third-party cloud for enterprise compliance. Codestral covers code, Mathstral covers math, Pixtral covers vision. Configurable reasoning effort in Small 4. Few LLM families offer this range.","Mixtral MoE models punch well above their weight in inference efficiency. Mistral Large 3 is competitive with frontier models on benchmarks. However, on raw capability benchmarks the top Mistral models are generally tier-2 behind GPT-4o and Claude Opus. The efficiency advantage is real and often decisive for cost-sensitive deployments.","Mistral 7B and Mixtral 8x7B were among the most downloaded open-weight models of 2023-2024. The lab is widely recognised as Europe's premier LLM provider, has $14B valuation, and models are available on every major cloud. Strong in enterprise and research circles.","Open-weight models are maximally portable — download once, run anywhere, no ongoing API dependency. Apache 2.0 permits commercial use and modification. Proprietary models (Mistral Large, Pixtral Large) carry some lock-in, but the API is OpenAI-compatible so switching costs are low.",{"tool_id":444,"name":445,"slug":446,"tooltip_description":447,"logo_url":448,"logo_bg":115,"pricing_model":449,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":450,"subcategory":451,"categories":452,"subcategories":454,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":456,"aliases":460,"score_reasonings":461,"published_date":337,"last_updated_date":107},158,"Google Gemma","google-gemma","Google DeepMind's family of open-weight LLMs, from compact edge models to a 31B dense flagship and a 26B MoE, built for self-hosting, on-device use, and fine-tuning under Apache 2.0.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgoogle-gemma.svg",{"slug":151,"display_name":85,"description":152},{"category_id":384,"name":374,"slug":385},{"subcategory_id":387,"name":388,"slug":389},[453],{"category_id":384,"name":374,"slug":385,"is_primary":3,"display_order":78},[455],{"subcategory_id":387,"name":388,"slug":389,"category_id":384,"is_primary":3,"display_order":78},[457,458,459],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":462,"flexibility":463,"performance":464,"popularity":465,"portability":466},"Running Gemma locally is accessible to any developer: `ollama run gemma3` requires no account, no API key, and no GPU on smaller models. The Hugging Face Transformers integration is industry-standard. Fine-tuning via LoRA or full fine-tuning adds moderate complexity but is well-documented.","Multiple sizes (1B to 31B), multiple architectures (dense and MoE), multimodal variants (PaliGemma), code-specialised variants (CodeGemma), domain fine-tunes (MedGemma), and embedding models (EmbeddingGemma). Deployable on any hardware — phones, laptops, consumer GPUs, cloud VMs. Runs with Ollama, vLLM, llama.cpp, JAX, PyTorch, Keras, and Transformers. Apache 2.0 (Gemma 4) allows full fine-tuning and redistribution.","Gemma models consistently top their weight class on standard benchmarks. Gemma 2 27B rivalled 70B-class models at release; Gemma 3 27B set new small-model records on MMLU and reasoning. Gemma 4 31B reaches 85.2% MMLU Pro and 89.2% AIME 2026. However, top-of-market reasoning (GPT-5, Claude Opus, Gemini 3.1 Pro) still exceeds the 31B ceiling.","The google-deepmind\u002Fgemma GitHub repository has ~5.1K stars; Gemma models on Hugging Face have millions of downloads. Widely used in on-device AI, edge deployment, and fine-tuning research. Strong adoption in the open-source ML community, though Meta Llama commands a larger absolute user base and mindshare.","Maximum portability — Gemma models run on CPUs, consumer GPUs, server GPUs, TPUs, mobile phones (E2B\u002FE4B variants), and Raspberry Pi-class hardware. GGUF quantisation via llama.cpp enables deployment anywhere. No cloud lock-in; weights are self-contained.",{"tool_id":468,"name":469,"slug":470,"tooltip_description":471,"logo_url":472,"logo_bg":115,"pricing_model":473,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":474,"subcategory":475,"categories":476,"subcategories":478,"flexibility_score":43,"performance_score":43,"portability_score":43,"is_featured":81,"tags":480,"aliases":484,"score_reasonings":485,"published_date":337,"last_updated_date":107},161,"DeepSeek","deepseek","Chinese open-weight LLM family under the MIT license, led by DeepSeek-V4.1-Flash, offering frontier-level results at very low API prices, with full self-hosting through vLLM, SGLang, and Ollama.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fdeepseek.svg",{"slug":62,"display_name":63,"description":64},{"category_id":384,"name":374,"slug":385},{"subcategory_id":387,"name":388,"slug":389},[477],{"category_id":384,"name":374,"slug":385,"is_primary":3,"display_order":78},[479],{"subcategory_id":387,"name":388,"slug":389,"category_id":384,"is_primary":3,"display_order":78},[481,482,483],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":486,"popularity":487,"portability":488,"flexibility":489,"performance":490},"The DeepSeek API (OpenAI-compatible) is trivial to call for developers already using OpenAI. Running distilled models via Ollama is also beginner-friendly. The steep part is self-hosting the full 671B V3\u002FR1 models — that requires multi-GPU infrastructure, tensor parallelism configuration in vLLM, and knowledge of quantisation trade-offs.","One of the fastest open-source repository growth stories in GitHub history. DeepSeek-R1 reached ~92K stars within weeks; total deepseek-ai org stars exceeded 170K by end of 2025. Massive developer adoption, significant industry impact, and strong Hugging Face download counts.","MIT license, Hugging Face weights, Ollama\u002FvLLM\u002FLM Studio support, and availability on Amazon Bedrock, Fireworks, Together AI, and OpenRouter. No lock-in whatsoever.","Open MIT weights with no use restrictions, publishable on any infrastructure, fine-tuneable with standard tools, and accessible via multiple APIs (direct, Bedrock, Together AI, Fireworks). Maximum possible flexibility for an LLM.","DeepSeek-V3 and R1 match or exceed GPT-4o on coding (HumanEval, SWE-bench), math (AIME, MATH-500), and reasoning benchmarks. R1 matches OpenAI o1. V4-Pro (2026) rivals the world's top closed models. Exceptional especially on coding tasks.",{"qwen":492,"mistral":493,"google-gemma":494,"deepseek":495},"Qwen is the default here. The Qwen 3.6 coding build at 27B is an 18 GB download and Qwen3-Coder at 30B is 19 GB, both with a 256K context limit, sized for a 24 GB graphics card or a 32 GB Mac. OpenCode's docs name the Qwen-Coder family as the one to try when local tool calls misbehave. Most releases are Apache 2.0, and every agent in this stack can run them.","Mistral's Devstral models are built for agentic coding. Devstral Small 2, a 24B model, is a 15 GB download with a context limit of 384K tokens, which leaves more room on a 24 GB card than the Qwen coders do. The weights are Apache 2.0 and come from a European lab. A strong second model to keep downloaded for comparison on your own repository.","Google's Gemma 4 is a general model that also codes. Its 26B size is a 16 GB download and 31B is 19 GB, with a 256K context limit, tool calling, and image input, which helps when a task includes a screenshot or a diagram. The 12B model at about 8 GB is the fallback for a graphics card with 12 or 16 GB, with shorter and simpler tasks in mind.","The DeepSeek models in Ollama's library that fit one graphics card are older ones: DeepSeek-Coder V2 at 16B is a 9 GB download, and the distilled DeepSeek-R1 sizes add a visible reasoning step. They suit cards with 12 to 16 GB. DeepSeek also publishes the weights of its current models under the MIT licence, with quantized builds from the community, so check a build's size against your memory before downloading it.",{},{"kind":498,"slug":385,"name":374,"href":499},"category","\u002Ftools\u002Fcategories\u002Fllm",{"name":501,"tools":502,"descriptions":558,"aliases":561,"see_all":562},"Version Control",[503,534],{"tool_id":504,"name":505,"slug":506,"tooltip_description":507,"logo_url":508,"logo_bg":60,"pricing_model":509,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":510,"subcategory":511,"categories":514,"subcategories":516,"flexibility_score":38,"performance_score":153,"portability_score":46,"is_featured":81,"tags":518,"aliases":525,"score_reasonings":528,"published_date":337,"last_updated_date":107},151,"GitHub","github","The largest Git hosting platform, where well over 100 million developers collaborate, review code, and manage projects through pull requests, issues, and a large integration ecosystem.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgithub.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":512,"name":501,"slug":513},50,"version-control",[515],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[517],{"subcategory_id":512,"name":501,"slug":513,"category_id":235,"is_primary":3,"display_order":78},[519,523,524],{"tag_id":520,"name":521,"slug":522,"tag_type":130},36,"CI\u002FCD","ci-cd",{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[526],{"alias":527,"is_primary":3},"GH",{"learning_curve":529,"flexibility":530,"performance":531,"popularity":532,"portability":533},"GitHub's UI is intuitive, but mastering Git concepts (rebasing, merge conflicts, branching strategies) and collaborative pull request workflows takes meaningful time — rated moderate difficulty.","Highly customisable branch protection, review policies, webhooks, and an extensive API\u002Fapp ecosystem. Some advanced governance features are gated behind paid tiers.","Generally reliable with 99.9% SLA for Enterprise, but the platform experienced capacity-related reliability incidents in 2025–2026 driven by AI agent traffic, pulling uptime below target.","Dominant market leader with ~70% SCM market share, 180M+ registered developers, 92% adoption among Fortune 100 companies. Effectively the industry default.","High vendor lock-in: GitHub Actions workflows, Projects, Packages, and Advanced Security are proprietary. Migrating a Git repo is easy; migrating the full workflow is not.",{"tool_id":535,"name":536,"slug":537,"tooltip_description":538,"logo_url":539,"logo_bg":60,"pricing_model":540,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":541,"subcategory":542,"categories":543,"subcategories":545,"flexibility_score":43,"performance_score":153,"portability_score":43,"is_featured":81,"tags":547,"aliases":551,"score_reasonings":552,"published_date":337,"last_updated_date":107},152,"GitLab","gitlab","An open-source, all-in-one DevSecOps platform offering Git hosting, code review, issue tracking, and built-in CI\u002FCD — available as a SaaS service or fully self-hosted.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgitlab.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":512,"name":501,"slug":513},[544],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[546],{"subcategory_id":512,"name":501,"slug":513,"category_id":235,"is_primary":3,"display_order":78},[548,549,550],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":553,"flexibility":554,"performance":555,"popularity":556,"portability":557},"Feature breadth is GitLab's biggest onboarding challenge — new users face pipelines, boards, security dashboards, and compliance tools all at once. Rated steep.","Highly flexible: customisable CI\u002FCD pipelines, branch protection, merge request approvals, REST\u002FGraphQL APIs, and full self-hosting. One of the most configurable platforms in its class.","Cloud performance is solid but not best-in-class; complex pipelines and large self-hosted deployments can lag. Rated average for its category.","16% SCM market share — firmly the second-largest platform globally, Gartner Magic Quadrant leader for DevOps, but well behind GitHub's ~70% dominance.","Open-source MIT licence with fully self-hostable CE and FOSS mirror. No vendor lock-in — highest portability score in its category.",{"github":559,"gitlab":560},"GitHub is the default here: the largest pull-request ecosystem, and where AI code reviewers and CI run. Keeping the model local does not keep the repository local, so code pushed to GitHub is on GitHub's servers. For most teams that is the existing arrangement and only the model stays private.","GitLab bundles source control, merge requests, and CI\u002FCD in one platform you can host yourself, the pick when the reason for a local model is that code must not leave your infrastructure at all. The agents only need git for local work, so everyday sessions are the same on either host.",{},{"kind":222,"slug":513,"name":501,"href":563},"\u002Ftools\u002Fcategories\u002Fdevelopment-tools\u002Fversion-control",{"server":565,"remote_access":646,"terminal":681,"code_review":810,"model_aggregator":894,"ci_cd":957},{"name":566,"tools":567,"descriptions":639,"aliases":642,"preface":643,"see_all":644},"Server",[568,597],{"tool_id":387,"name":569,"slug":570,"tooltip_description":571,"logo_url":572,"logo_bg":60,"pricing_model":573,"learning_curve_score":46,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":574,"subcategory":577,"categories":581,"subcategories":583,"flexibility_score":153,"performance_score":38,"portability_score":38,"is_featured":81,"tags":585,"aliases":590,"score_reasonings":591,"published_date":337,"last_updated_date":107},"Hetzner","hetzner","German cloud and dedicated server provider offering VPS, bare-metal servers, and storage at prices far below the big clouds, with data centers in the EU, the US, and Singapore.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fhetzner.svg",{"slug":344,"display_name":345,"description":346},{"category_id":43,"name":575,"slug":576},"Hosting & Cloud","hosting-cloud",{"subcategory_id":578,"name":579,"slug":580},42,"VPS & Servers","vps-and-servers",[582],{"category_id":43,"name":575,"slug":576,"is_primary":3,"display_order":78},[584],{"subcategory_id":578,"name":579,"slug":580,"category_id":43,"is_primary":3,"display_order":78},[586,587],{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":69,"name":588,"slug":589,"tag_type":87},"Multi-region","multi-region",[],{"learning_curve":592,"flexibility":593,"performance":594,"portability":595,"popularity":596},"Standard VPS setup; any developer comfortable with Linux can deploy in an afternoon.","Standard VPS; configure anything on Linux; no managed abstraction layer constraining choices.","Competitive hardware at low cost; dedicated and VPS servers deliver strong baseline performance.","Standard Linux VPS; highly portable with no managed-service abstraction to escape.","Popular among European developers and cost-conscious teams; strong reputation for value.",{"tool_id":598,"name":599,"slug":600,"tooltip_description":601,"logo_url":602,"logo_bg":60,"pricing_model":603,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":607,"subcategory":608,"categories":612,"subcategories":614,"flexibility_score":38,"performance_score":38,"portability_score":38,"is_featured":81,"tags":616,"aliases":632,"score_reasonings":633,"published_date":337,"last_updated_date":107},166,"DigitalOcean","digitalocean","Developer-focused cloud platform offering VPS (Droplets), managed Kubernetes, App Platform PaaS, and managed databases. Known for simple pricing and excellent documentation.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fdigitalocean.svg",{"slug":604,"display_name":605,"description":606},"usage_based","Usage-Based","Pricing scales with consumption: API calls, data volume, compute time, or similar metered units.",{"category_id":43,"name":575,"slug":576},{"subcategory_id":609,"name":610,"slug":611},41,"App Hosting","app-hosting",[613],{"category_id":43,"name":575,"slug":576,"is_primary":3,"display_order":78},[615],{"subcategory_id":609,"name":610,"slug":611,"category_id":43,"is_primary":3,"display_order":78},[617,618,619,623,624,628],{"tag_id":323,"name":324,"slug":325,"tag_type":135},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":620,"name":621,"slug":622,"tag_type":87},20,"Auto-scaling","auto-scaling",{"tag_id":69,"name":588,"slug":589,"tag_type":87},{"tag_id":625,"name":626,"slug":627,"tag_type":87},24,"Docker Compatible","docker-compatible",{"tag_id":629,"name":630,"slug":631,"tag_type":130},28,"Web Development","web-development",[],{"learning_curve":634,"flexibility":635,"performance":636,"popularity":637,"portability":638},"DigitalOcean is specifically designed for developers rather than enterprise ops teams. The control panel is clean, Droplets launch in under 60 seconds, and App Platform requires no infrastructure knowledge at all. Thousands of community tutorials lower the barrier for beginners.","Offers VPS, PaaS, and managed Kubernetes under one roof, covering most deployment patterns from simple static sites to complex microservices. Custom networking, firewalls, VPC peering, and a full API give teams significant control. Missing serverless functions and some niche managed services limit the ceiling slightly.","NVMe SSDs and dedicated vCPU Droplets deliver competitive single-tenant performance. The 99.99% uptime SLA holds in practice per benchmark studies. Lacks the global edge network density of Cloudflare or the ultra-low-latency primitives of AWS.","Consistently ranks in the top 5 cloud providers for developer and startup usage. Strong presence on StackOverflow, Reddit, and developer surveys. Over 600,000 active customers as of 2024. Widely used as a first production cloud by self-taught developers.","Standard Linux VMs and Docker\u002FKubernetes workloads migrate freely to any other cloud or bare metal. No proprietary runtime lock-in. Spaces is S3-compatible. App Platform lock-in is higher — migrating away requires containerizing or adapting the build pipeline.",{"hetzner":640,"digitalocean":641},"Add a Hetzner GPU server when no machine at home can hold the model: the entry dedicated server has a 24 GB NVIDIA card and 64 GB of RAM for about €214 a month plus a one-time setup fee, enough for the 15 to 19 GB coding models. The model then answers around the clock, on hardware that is rented and not local.","Add a DigitalOcean GPU Droplet for occasional sessions: the smallest has a 20 GB NVIDIA card at $0.76 an hour, about $550 if left on for a month, so it pays to destroy it after work and keep the models on a volume. A fit for trying larger models before buying a graphics card.",{},"Add a server when you want the agent to keep running somewhere other than your own laptop, reachable at any time and not tied to your machine staying on. Some agents also offer their own managed cloud sessions as an alternative to self-hosting; check the stack's own description for details.",{"kind":498,"slug":576,"name":575,"href":645},"\u002Ftools\u002Fcategories\u002Fhosting-cloud",{"name":647,"tools":648,"descriptions":677,"aliases":679,"preface":680,"see_all":35},"Remote Access",[649],{"tool_id":650,"name":651,"slug":652,"tooltip_description":653,"logo_url":654,"logo_bg":60,"pricing_model":655,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":656,"subcategory":659,"categories":663,"subcategories":665,"flexibility_score":38,"performance_score":38,"portability_score":153,"is_featured":81,"tags":667,"aliases":670,"score_reasonings":671,"published_date":107,"last_updated_date":35},275,"Tailscale","tailscale","Tailscale is a zero-config mesh VPN built on WireGuard that connects your devices and servers into one private network, reachable from anywhere without port forwarding.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Ftailscale.svg",{"slug":62,"display_name":63,"description":64},{"category_id":89,"name":657,"slug":658},"DevOps & CI\u002FCD","devops-cicd",{"subcategory_id":660,"name":661,"slug":662},70,"Tunneling & Secure Access","tunneling-secure-access",[664],{"category_id":89,"name":657,"slug":658,"is_primary":3,"display_order":78},[666],{"subcategory_id":660,"name":661,"slug":662,"category_id":89,"is_primary":3,"display_order":78},[668,669],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":672,"flexibility":673,"performance":674,"popularity":675,"portability":676},"Installing a client and authenticating via SSO puts a device on the tailnet in minutes, with MagicDNS and default ACLs handling the rest — no manual key exchange, port forwarding, or router configuration the way a raw WireGuard or OpenVPN setup requires.","Subnet routers, exit nodes, Tailscale SSH, ACL tags, and Funnel cover use cases well beyond simple point-to-point connectivity, and Headscale offers a fully self-hosted control-plane path for teams that want it.","Connections negotiate directly between peers over WireGuard whenever NAT traversal succeeds, falling back to DERP relays only when a direct path isn't reachable, which keeps overhead close to raw WireGuard rather than routing everything through a central VPN server.","A name that comes up constantly in engineering blogs, self-hosting communities, and general developer conversation about remote access — recognition extends well past the networking and DevOps specialists who actually configure it, unlike the more narrowly-known tunnel tools it's sometimes compared against.","Clients are open source and cover every major desktop, mobile, and server platform, but the default coordination server is Tailscale's own hosted service; Headscale exists specifically so self-hosters can replace that layer without losing client compatibility, though it isn't the default path.",{"tailscale":678},"Add Tailscale to reach the model machine from a laptop or phone anywhere: both join a private network, and the agent points at the desktop's or the server's address as if it were local, with nothing exposed to the internet. LM Studio's own LM Link is built on the same kind of mesh. The personal tier is free.",{},"Add remote access when you want to reach an agent running on another machine — a VPS or a home server — without exposing it to anyone but you.",{"name":682,"tools":683,"descriptions":802,"aliases":808,"preface":809,"see_all":35},"Terminal",[684,710,733,756,779],{"tool_id":685,"name":686,"slug":687,"tooltip_description":688,"logo_url":689,"logo_bg":115,"pricing_model":690,"learning_curve_score":46,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":691,"subcategory":692,"categories":696,"subcategories":698,"flexibility_score":38,"performance_score":43,"portability_score":153,"is_featured":81,"tags":700,"aliases":703,"score_reasonings":704,"published_date":107,"last_updated_date":35},279,"Ghostty","ghostty","A GPU-accelerated terminal emulator built by Mitchell Hashimoto that renders with native UI components on each platform instead of a cross-platform toolkit.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fghostty.svg",{"slug":151,"display_name":85,"description":152},{"category_id":235,"name":236,"slug":237},{"subcategory_id":693,"name":694,"slug":695},72,"Terminals","terminals",[697],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[699],{"subcategory_id":693,"name":694,"slug":695,"category_id":235,"is_primary":3,"display_order":78},[701,702],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":132,"name":133,"slug":134,"tag_type":135},[],{"learning_curve":705,"flexibility":706,"performance":707,"popularity":708,"portability":709},"Ships with sensible defaults and native settings UI on macOS, so most users are productive immediately; deeper customization (shaders, keybindings) means editing the plain-text config file.","Plain-text config covers fonts, colors, keybindings, and window behavior, and custom GLSL shaders allow visual customization beyond what most terminal emulators expose, though there's no plugin\u002Fextension marketplace.","GPU-accelerated native rendering plus a stated focus on correct emulation of demanding TUI applications puts it among the fastest and most reliable terminals in independent comparisons.","Grew to 60K+ GitHub stars in under two years and generated significant coverage in developer press, largely on the strength of Mitchell Hashimoto's reputation, but that recognition is still concentrated among developers who follow terminal-customization and HashiCorp-adjacent communities rather than the broader developer population.","Runs natively on macOS and Linux with a portable plain-text config, but has no native Windows build — the underlying parsing library targets Windows as future groundwork, not a shipping GUI today.",{"tool_id":711,"name":712,"slug":713,"tooltip_description":714,"logo_url":715,"logo_bg":115,"pricing_model":716,"learning_curve_score":153,"popularity_score":46,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":717,"subcategory":718,"categories":719,"subcategories":721,"flexibility_score":153,"performance_score":43,"portability_score":43,"is_featured":81,"tags":723,"aliases":726,"score_reasonings":727,"published_date":107,"last_updated_date":35},280,"Alacritty","alacritty","A GPU-accelerated, cross-platform terminal emulator focused on raw rendering speed and minimal, text-based configuration.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Falacritty.svg",{"slug":151,"display_name":85,"description":152},{"category_id":235,"name":236,"slug":237},{"subcategory_id":693,"name":694,"slug":695},[720],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[722],{"subcategory_id":693,"name":694,"slug":695,"category_id":235,"is_primary":3,"display_order":78},[724,725],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":132,"name":133,"slug":134,"tag_type":135},[],{"learning_curve":728,"flexibility":729,"performance":730,"popularity":731,"portability":732},"Works immediately with sane defaults, but real customization means hand-editing a TOML file, and tabs\u002Fsplits only appear once it's paired with a multiplexer like tmux — a step GUI-configured terminals don't require.","Every color, font, and keybinding is configurable via plain text, and an IPC socket allows scripted control, but there's no plugin\u002Fextension system — functionality beyond rendering is delegated entirely to external tools.","GPU-accelerated OpenGL rendering consistently benchmarks among the lowest-latency, highest-throughput terminal emulators available, even under heavy scrollback.","Well known among developers who actively customize their terminal setup, with 65K+ GitHub stars, but recognition is concentrated in that power-user segment rather than the broader developer population, most of whom use their OS default or an IDE's integrated terminal.","Runs natively on Linux, macOS, Windows, and BSD from one codebase, with a plain-text config that ports across machines unchanged and no dependency on any single vendor or platform.",{"tool_id":734,"name":735,"slug":736,"tooltip_description":737,"logo_url":738,"logo_bg":115,"pricing_model":739,"learning_curve_score":153,"popularity_score":46,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":740,"subcategory":741,"categories":742,"subcategories":744,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":746,"aliases":749,"score_reasonings":750,"published_date":107,"last_updated_date":35},281,"WezTerm","wezterm","A GPU-accelerated terminal emulator with a built-in multiplexer and a full Lua scripting environment for configuration.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fwezterm.svg",{"slug":151,"display_name":85,"description":152},{"category_id":235,"name":236,"slug":237},{"subcategory_id":693,"name":694,"slug":695},[743],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[745],{"subcategory_id":693,"name":694,"slug":695,"category_id":235,"is_primary":3,"display_order":78},[747,748],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":132,"name":133,"slug":134,"tag_type":135},[],{"learning_curve":751,"flexibility":752,"performance":753,"popularity":754,"portability":755},"Ships with usable defaults, but unlocking its real power means writing Lua rather than editing a flat config file, which asks more of users with no scripting background than simpler emulators do.","Configuration is a full Lua environment supporting conditional logic and event hooks, on top of a built-in multiplexer and native SSH client — capabilities most terminal emulators leave to external tools entirely.","GPU-accelerated Rust rendering keeps it fast and responsive, though the built-in multiplexer and session-persistence layer add overhead that minimal, rendering-only terminals don't carry.","Recognized within the terminal-customization and cross-platform-dev communities, particularly for its Windows support, but that recognition stays inside the power-user niche rather than reaching developers who don't customize their terminal setup.","The same Lua config runs unmodified on macOS, Linux, Windows, and FreeBSD, and SSH-backed sessions can be reattached from any of them, with no vendor tie-in of any kind.",{"tool_id":757,"name":758,"slug":759,"tooltip_description":760,"logo_url":761,"logo_bg":115,"pricing_model":762,"learning_curve_score":153,"popularity_score":46,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":763,"subcategory":764,"categories":765,"subcategories":767,"flexibility_score":43,"performance_score":43,"portability_score":153,"is_featured":81,"tags":769,"aliases":772,"score_reasonings":773,"published_date":107,"last_updated_date":35},282,"Kitty","kitty","A GPU-accelerated terminal emulator with a Python-based plugin system (kittens), built-in image display, and its own protocol extensions.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fkitty.png",{"slug":151,"display_name":85,"description":152},{"category_id":235,"name":236,"slug":237},{"subcategory_id":693,"name":694,"slug":695},[766],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[768],{"subcategory_id":693,"name":694,"slug":695,"category_id":235,"is_primary":3,"display_order":78},[770,771],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":132,"name":133,"slug":134,"tag_type":135},[],{"learning_curve":774,"flexibility":775,"performance":776,"popularity":777,"portability":778},"Basic usage works out of the box, but its deeper capabilities — writing kittens, using the remote-control API — require Python scripting knowledge that casual users won't already have.","A Python-based plugin system (kittens), a full remote-control API, and built-in tabs\u002Fsplits\u002Flayouts give it more programmable surface area than emulators that only expose static configuration.","SIMD-accelerated, multithreaded GPU rendering keeps it competitive with the fastest terminal emulators available, and it originated protocol extensions other tools later adopted for their own performance gains.","Well known and long-established among terminal power users, and its protocol contributions are recognized industry-wide among emulator authors, but general developer awareness of Kitty specifically stays inside that niche.","Runs natively on Linux, macOS, and BSD, but Windows users are limited to running it inside WSL rather than a native build, which is a real gap relative to fully cross-platform alternatives.",{"tool_id":780,"name":781,"slug":782,"tooltip_description":783,"logo_url":784,"logo_bg":60,"pricing_model":785,"learning_curve_score":46,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":786,"subcategory":787,"categories":788,"subcategories":790,"flexibility_score":38,"performance_score":153,"portability_score":65,"is_featured":81,"tags":792,"aliases":795,"score_reasonings":796,"published_date":107,"last_updated_date":35},283,"iTerm2","iterm2","A free, open-source terminal replacement for macOS with a deep feature set built up over more than a decade of development.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fiterm2.svg",{"slug":151,"display_name":85,"description":152},{"category_id":235,"name":236,"slug":237},{"subcategory_id":693,"name":694,"slug":695},[789],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[791],{"subcategory_id":693,"name":694,"slug":695,"category_id":235,"is_primary":3,"display_order":78},[793,794],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":797,"flexibility":798,"performance":799,"popularity":800,"portability":801},"Ships with a full native macOS preferences GUI and sensible defaults, so day-one usage is straightforward; the depth of options (triggers, profiles, the Python scripting API) takes longer to fully explore than a minimal terminal's config file.","Profiles, triggers, a status bar with pluggable components, and a Python API for scripting give it one of the broader customization surfaces among terminal emulators, short of a full plugin marketplace.","Handles everyday terminal workloads well but isn't GPU-accelerated the way the newer Rust\u002FZig-based emulators are, and has a longstanding reputation for slowing down on very large output streams compared to that group.","Predates GitHub-stars-driven discovery and built its install base as the default non-default macOS terminal for over a decade through Homebrew and direct downloads rather than repository traffic — broadly recognized among Mac developers specifically, but that recognition doesn't extend past macOS the way a cross-platform tool's would.","macOS-only by design with no Windows or Linux build on the roadmap — the least portable of the terminal emulators in this catalog.",{"ghostty":803,"alacritty":804,"wezterm":805,"kitty":806,"iterm2":807},"A GPU-accelerated terminal with native macOS and Linux integration, one of the terminals OpenCode's docs name, and a solid default for keeping an agent and a runtime's logs open side by side.","A minimal, GPU-accelerated terminal that does little beyond drawing text fast; pick it when a multiplexer or window manager already handles tabs and splits, and every bit of graphics memory should go to the model.","A GPU-accelerated terminal with built-in tabs, splits, and multiplexing configured in Lua, able to stand in for both the terminal and tmux, with the agent in one pane and the model server in another.","A GPU-accelerated terminal with its own graphics protocol and a scripting layer called kittens, for developers who want deep customization of the window their terminal agent lives in.","The long-standing free macOS terminal, with split panes, profiles, and triggers. The familiar pick on a Mac, where Apple Silicon is also the easiest hardware to run a local model on.",{},"Add a terminal when you want a faster, more configurable place to run the agent than your OS default — most agent CLIs live here all day.",{"name":811,"tools":812,"descriptions":888,"aliases":892,"preface":893,"see_all":35},"Code Review",[813,840,865],{"tool_id":814,"name":815,"slug":816,"tooltip_description":817,"logo_url":818,"logo_bg":115,"pricing_model":819,"learning_curve_score":65,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":820,"subcategory":821,"categories":825,"subcategories":827,"flexibility_score":38,"performance_score":153,"portability_score":153,"is_featured":81,"tags":829,"aliases":833,"score_reasonings":834,"published_date":107,"last_updated_date":35},284,"CodeRabbit","coderabbit","AI code reviewer that comments on every pull request in GitHub, GitLab, Bitbucket, and Azure DevOps, combining language-model review with more than 40 static analyzers. The same reviews also run in IDE extensions and a CLI.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fcoderabbit.svg",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":822,"name":823,"slug":824},73,"AI Code Review","ai-code-review",[826],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[828],{"subcategory_id":822,"name":823,"slug":824,"category_id":235,"is_primary":3,"display_order":78},[830,831,832],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"flexibility":835,"learning_curve":836,"performance":837,"popularity":838,"portability":839},"Path-specific instructions, learnings from PR replies, custom pre-merge checks, and four Git hosts plus IDE and CLI surfaces give teams a lot of control over what gets reviewed and how. The underlying models are chosen by CodeRabbit, not the user.","Setup is installing a Git app and picking repositories; the first review arrives on the next pull request with no configuration. The YAML config and path instructions are optional refinements.","Reviews arrive within minutes and PR summaries are widely praised, but independent bug-catch benchmarks put its detection rate well behind Greptile and some other reviewers, and noisy comments are a common complaint.","The most-installed AI app on the GitHub Marketplace, with over two million connected repositories, more than 17,000 customers, and a $1.5B valuation. Well known among developers who work with pull requests, though AI code review is still a young category many developers have not adopted.","Works across GitHub, GitLab, Bitbucket, and Azure DevOps and sits outside the codebase, so switching reviewers is an app uninstall. It is a hosted SaaS, and learnings and configuration do not carry over to other tools.",{"tool_id":841,"name":842,"slug":843,"tooltip_description":844,"logo_url":845,"logo_bg":115,"pricing_model":846,"learning_curve_score":46,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":847,"subcategory":848,"categories":849,"subcategories":851,"flexibility_score":38,"performance_score":38,"portability_score":38,"is_featured":81,"tags":853,"aliases":858,"score_reasonings":859,"published_date":107,"last_updated_date":35},285,"Greptile","greptile","AI code reviewer that indexes the whole repository into a code graph and uses it to review pull requests, catching bugs that span several files. Supports GitHub, GitLab, and Bitbucket, with self-hosting on Enterprise.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgreptile.png",{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":822,"name":823,"slug":824},[850],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[852],{"subcategory_id":822,"name":823,"slug":824,"category_id":235,"is_primary":3,"display_order":78},[854,855,856,857],{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":860,"flexibility":861,"performance":862,"popularity":863,"portability":864},"Connecting a Git host and indexing a repository is quick, but getting the most out of it means writing custom rules and understanding the credit model and when a TREX review is worth three credits.","Plain-language custom rules, a choice of standard or execution-backed reviews, an API, and on Enterprise a bring-your-own-model option give teams real control over how reviews run.","Independent comparisons put its raw bug-catch rate at the top of the mainstream AI reviewers, driven by the codebase graph, and TREX adds execution evidence that static reading cannot produce.","Well regarded among teams evaluating AI code review, with more than 2,000 customers and Benchmark backing, but far less widely known than CodeRabbit and largely unknown to developers outside that decision.","Supports GitHub, GitLab, and Bitbucket in the cloud and can be self-hosted or run air-gapped with a team's own models on Enterprise, which few rivals offer. Azure DevOps is the notable gap.",{"tool_id":866,"name":867,"slug":868,"tooltip_description":869,"logo_url":870,"logo_bg":115,"pricing_model":871,"learning_curve_score":65,"popularity_score":153,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":872,"subcategory":873,"categories":874,"subcategories":876,"flexibility_score":153,"performance_score":38,"portability_score":46,"is_featured":81,"tags":878,"aliases":881,"score_reasonings":882,"published_date":107,"last_updated_date":35},286,"Cursor Bugbot","cursor-bugbot","Cursor's AI pull request reviewer, which runs as a GitHub or GitLab app and flags logic bugs, edge cases, and security issues on each PR. It works on any repository, whether or not the code was written in Cursor.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fcursor-bugbot.svg",{"slug":604,"display_name":605,"description":606},{"category_id":235,"name":236,"slug":237},{"subcategory_id":822,"name":823,"slug":824},[875],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[877],{"subcategory_id":822,"name":823,"slug":824,"category_id":235,"is_primary":3,"display_order":78},[879,880],{"tag_id":97,"name":98,"slug":99,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":883,"flexibility":884,"performance":885,"popularity":886,"portability":887},"Enabling it is a toggle in the Cursor dashboard plus a Git app install, and it reviews the next pull request with no rules written. BUGBOT.md files are optional context.","Nested BUGBOT.md rules, team rules, and effort levels let teams steer what it looks for, but the models and review style are Cursor's choice and it only covers GitHub and GitLab.","Reviews run automatically on each update and got about three times faster in a mid-2026 update; Cursor reports developers resolving around 80% of flagged bugs by merge, with the high effort setting finding more on complex changes.","The Cursor name is widely known and Bugbot reviews more than two million pull requests a month, but the reviewer itself is known mainly to Cursor users and teams comparing AI code review tools.","Works on any repository on GitHub or GitLab, but setup and billing are tied to a Cursor account, rules live in Cursor-specific BUGBOT.md files, and Autofix depends on Cursor's cloud agents.",{"coderabbit":889,"greptile":890,"cursor-bugbot":891},"Add CodeRabbit when you want a second opinion on what a local model wrote: it installs as a GitHub or GitLab app, summarizes each pull request, and leaves line-level comments before anyone merges. It is a hosted service, so the diff is sent to it, which matters if the model stayed local for privacy.","Add Greptile when review should read the whole repository and not just the diff: it indexes the codebase so its pull request comments carry wider context. It is a hosted service by default, and its Enterprise plan adds the option to self-host it on your own infrastructure, which suits a setup built to keep code in house.","Add Cursor Bugbot when a bug-focused reviewer is enough: it reviews GitHub or GitLab pull requests whatever wrote the code, and bills through a Cursor account and plan. A hosted service, like CodeRabbit.",{},"Add code review when you want an AI reading every pull request before it merges: it comments on the diff so bugs, security issues, and inconsistencies surface before a human has to catch them.",{"name":895,"tools":896,"descriptions":950,"aliases":953,"preface":954,"see_all":955},"Model Aggregator",[897,924],{"tool_id":898,"name":899,"slug":900,"tooltip_description":901,"logo_url":902,"logo_bg":115,"pricing_model":903,"learning_curve_score":65,"popularity_score":38,"hosting_assignment_type":904,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":905,"subcategory":906,"categories":910,"subcategories":912,"flexibility_score":38,"performance_score":38,"portability_score":153,"is_featured":81,"tags":914,"aliases":917,"score_reasonings":918,"published_date":107,"last_updated_date":143},218,"OpenRouter","openrouter","Unified, OpenAI-compatible API to 500+ models across 80+ providers, with automatic provider fallback and one credits-based bill: the hosted, nothing-to-deploy counterpart to a self-hosted gateway like LiteLLM.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fopenrouter.svg",{"slug":604,"display_name":605,"description":606},"managed_only",{"category_id":69,"name":70,"slug":71},{"subcategory_id":907,"name":908,"slug":909},63,"AI Model Aggregators","ai-model-aggregators",[911],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},[913],{"subcategory_id":907,"name":908,"slug":909,"category_id":69,"is_primary":3,"display_order":78},[915,916],{"tag_id":127,"name":128,"slug":129,"tag_type":130},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"flexibility":919,"performance":920,"portability":921,"popularity":922,"learning_curve":923},"500+ models across 80+ providers, custom data policies, and a Bring-Your-Own-Key mode cover most routing needs, though there's no control over the underlying routing logic the way a self-hosted gateway offers.","Edge-deployed routing and automatic provider fallback keep added latency low and reliability high, backed by real scale (200+ trillion tokens\u002Fmonth).","Closed-source, hosted-only SaaS with no self-hosting option, credits and API keys are OpenRouter-specific and don't transfer if switching to a self-hosted gateway.","Aggregates 400+ models from 60+ providers behind one OpenAI-style API and is usually the first stop for developers who want to try many models without juggling separate provider keys. A fixture of the LLM-app-building corner of AI development specifically, not a name outside that corner.","A single OpenAI-compatible API key and endpoint gives instant access to hundreds of models, no infrastructure setup or provider-specific integration work required.",{"tool_id":925,"name":926,"slug":927,"tooltip_description":928,"logo_url":929,"logo_bg":115,"pricing_model":930,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":66,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":931,"subcategory":932,"categories":933,"subcategories":935,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":937,"aliases":942,"score_reasonings":943,"published_date":107,"last_updated_date":949},221,"LiteLLM","litellm","The open-source AI gateway you run yourself: one OpenAI-compatible endpoint in front of more than 100 model providers, with virtual keys, budgets, and spend tracking. It is the self-hosted counterpart to OpenRouter.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Flitellm.png",{"slug":62,"display_name":63,"description":64},{"category_id":69,"name":70,"slug":71},{"subcategory_id":907,"name":908,"slug":909},[934],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},[936],{"subcategory_id":907,"name":908,"slug":909,"category_id":69,"is_primary":3,"display_order":78},[938,939,940,941],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":127,"name":128,"slug":129,"tag_type":130},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[],{"learning_curve":944,"flexibility":945,"performance":946,"popularity":947,"portability":948},"Requires standing up and maintaining real infrastructure, PostgreSQL, Redis, the proxy server itself, meaningfully more setup than a zero-ops hosted aggregator like OpenRouter.","Full control over routing logic, guardrails, virtual keys, and spend limits across 100+ providers, the most configurable option in the aggregator genre precisely because you own the deployment.","8ms P95 latency at 1,000 RPS and a Rust core deliver genuinely fast proxy performance, though real-world throughput depends on the self-hosted infrastructure backing it.","56K+ GitHub stars and production adoption at Stripe, Google ADK, Greptile, and OpenHands make it the clear self-hosted leader in this genre.","MIT licensed, genuinely self-hostable anywhere with Docker or the provided Terraform modules, no vendor lock-in to a hosted service.","2026-10-06",{"openrouter":951,"litellm":952},"Add OpenRouter as the way out for tasks a local model can't finish: one key for hosted models, including the larger siblings of the open models used here, selected in the agent like any other provider. Those requests do leave your machine, at provider prices plus 5.5% on credit purchases, so keep it for work that is allowed to.","Add LiteLLM when several developers share the setup: a self-hosted proxy that puts the local runtime and any hosted models behind one address, with a key and a budget per person. The agents connect to it as an OpenAI-compatible provider, and nothing leaves your network unless a hosted model is configured behind it. Free to self-host.",{},"Add a model aggregator when you want one API key and one bill for models from many providers, with automatic fallback when one of them is down, instead of setting up each provider separately.",{"kind":222,"slug":909,"name":908,"href":956},"\u002Ftools\u002Fcategories\u002Fai-infrastructure\u002Fai-model-aggregators",{"name":521,"tools":958,"descriptions":1022,"aliases":1025,"preface":1026,"see_all":1027},[959,995],{"tool_id":960,"name":961,"slug":962,"tooltip_description":963,"logo_url":964,"logo_bg":60,"pricing_model":965,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":966,"subcategory":967,"categories":971,"subcategories":973,"flexibility_score":43,"performance_score":38,"portability_score":38,"is_featured":81,"tags":975,"aliases":986,"score_reasonings":989,"published_date":337,"last_updated_date":107},164,"GitHub Actions","github-actions","GitHub's integrated CI\u002FCD platform that automates build, test, and deployment workflows using YAML-based configurations. Runs on GitHub-hosted or self-hosted runners with a rich marketplace of pre-built integrations.","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgithub-actions.svg",{"slug":62,"display_name":63,"description":64},{"category_id":89,"name":657,"slug":658},{"subcategory_id":968,"name":969,"slug":970},34,"CI\u002FCD Pipelines","cicd-pipelines",[972],{"category_id":89,"name":657,"slug":658,"is_primary":3,"display_order":78},[974],{"subcategory_id":968,"name":969,"slug":970,"category_id":89,"is_primary":3,"display_order":78},[976,977,978,979,980,981],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},{"tag_id":520,"name":521,"slug":522,"tag_type":130},{"tag_id":625,"name":626,"slug":627,"tag_type":87},{"tag_id":982,"name":983,"slug":984,"tag_type":985},45,"Declarative","declarative","paradigm",[987],{"alias":988,"is_primary":3},"GHA",{"flexibility":990,"learning_curve":991,"performance":992,"popularity":993,"portability":994},"Self-hosted runner support across any OS or cloud, custom labels, matrix builds, Kubernetes scaling via Actions Runner Controller, and a marketplace of thousands of community-built actions give teams virtually unlimited configuration options for any workflow or environment.","Basic single-job workflows are accessible to any developer comfortable with YAML, and GitHub's documentation lowers the barrier further. However, advanced patterns — composite actions, reusable workflows, OIDC-based cloud auth, and conditional matrix strategies — involve non-obvious syntax and a complex permission model that requires meaningful time to master.","The platform handles massive scale reliably, but GitHub-hosted runner performance can vary between runs, making consistent benchmarking difficult. Teams running on self-hosted or larger hosted runners achieve stable, high throughput; the managed offering trades predictable latency for zero infrastructure overhead.","GitHub Actions is the dominant CI\u002FCD platform with tens of millions of repositories using it, thousands of marketplace actions, and widespread enterprise adoption. It is cited as the most-used CI\u002FCD solution in multiple developer surveys.","Self-hosted runners can run on any cloud or on-premises infrastructure, and the YAML workflow model is readable and auditable. The main constraint is tight coupling to GitHub events and APIs — moving workflows to another CI\u002FCD platform requires meaningful rewriting rather than a simple lift-and-shift.",{"tool_id":996,"name":997,"slug":998,"tooltip_description":999,"logo_url":539,"logo_bg":60,"pricing_model":1000,"learning_curve_score":153,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":1001,"subcategory":1002,"categories":1003,"subcategories":1005,"flexibility_score":43,"performance_score":153,"portability_score":153,"is_featured":81,"tags":1007,"aliases":1015,"score_reasonings":1016,"published_date":337,"last_updated_date":257},165,"GitLab CI\u002FCD","gitlab-cicd","GitLab's built-in CI\u002FCD system configured through .gitlab-ci.yml files stored in your repository. Supports both GitLab-hosted and self-hosted runners for flexible pipeline execution across diverse environments.",{"slug":62,"display_name":63,"description":64},{"category_id":89,"name":657,"slug":658},{"subcategory_id":968,"name":969,"slug":970},[1004],{"category_id":89,"name":657,"slug":658,"is_primary":3,"display_order":78},[1006],{"subcategory_id":968,"name":969,"slug":970,"category_id":89,"is_primary":3,"display_order":78},[1008,1009,1010,1011,1012,1013,1014],{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},{"tag_id":520,"name":521,"slug":522,"tag_type":130},{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":625,"name":626,"slug":627,"tag_type":87},{"tag_id":982,"name":983,"slug":984,"tag_type":985},[],{"learning_curve":1017,"flexibility":1018,"performance":1019,"popularity":1020,"portability":1021},"The core .gitlab-ci.yml syntax is approachable for developers with YAML and Linux fundamentals, and GitLab's documentation is thorough. However, the platform's breadth — runner executors, pipeline inheritance, CI\u002FCD Catalog, merge trains, and security scanning configuration — creates a long tail of advanced concepts that teams discover gradually over months.","Multiple runner executor types (Docker, Kubernetes, Shell, machine), unlimited self-hosted runner capacity, parent-child pipelines, reusable CI\u002FCD Components, and the ability to self-host the entire platform give teams complete control over their pipeline environment and infrastructure.","Pipeline performance depends heavily on runner configuration and caching strategy. Shared GitLab-hosted runners can experience queue delays during peak periods, and Docker+machine executors add startup overhead. Teams running optimized self-hosted Kubernetes runners with effective caching achieve significantly faster pipelines, but the default shared experience is average for the CI\u002FCD category.","GitLab CI\u002FCD is used by over 100,000 organizations and is particularly strong among enterprises and security-conscious DevOps teams. It consistently ranks among the top CI\u002FCD platforms in developer surveys, though GitHub Actions has overtaken it in raw adoption volume — especially among open-source and smaller team segments.","GitLab CE is fully open source and self-hostable on any infrastructure, which is a strong portability advantage. However, .gitlab-ci.yml pipelines are tightly coupled to GitLab's API and runner ecosystem — migrating pipeline definitions to another CI\u002FCD system (GitHub Actions, CircleCI, etc.) requires substantial rewriting rather than a simple port.",{"github-actions":1023,"gitlab-cicd":1024},"Add GitHub Actions when the repository lives on GitHub and every push the agent makes should run the tests. An automatic test run is worth more here than with a hosted model, since a smaller model makes more mistakes that only the tests catch.","Add GitLab CI\u002FCD when the project is on GitLab and needs the same automated tests on every push, defined in a .gitlab-ci.yml file that can run entirely on your own runners, next to the self-hosted repository and the local model.",{},"Add CI\u002FCD when you want a dedicated pipeline for running tests, linting, or multi-stage builds before a deploy goes out. Many hosting platforms already redeploy automatically on every push on their own — a CI\u002FCD tool adds the most value on top of that by gating the deploy on a passing test suite, and matters even more when the hosting choice does not auto-deploy at all, such as a self-hosted server.",{"kind":222,"slug":970,"name":969,"href":1028},"\u002Ftools\u002Fcategories\u002Fdevops-cicd\u002Fcicd-pipelines",[1030,1040,1047],{"slug":1031,"name":1032,"description":1033,"icon_url":1034,"addition_types":1035},"developer_workflow","Developer workflow","Where the agent runs, how you reach it, and the tools you work in alongside it.","https:\u002F\u002Fassets.tekyous.dev\u002Ficons\u002Faddition-groups\u002Fdeveloper_workflow.svg",[1036,1037,1038,1039],"server","remote_access","terminal","code_review",{"slug":1041,"name":1042,"description":1043,"icon_url":1044,"addition_types":1045},"ai","AI","Models, agents, and the pieces around them: retrieval, sandboxes, and model access.","https:\u002F\u002Fassets.tekyous.dev\u002Ficons\u002Faddition-groups\u002Fai.svg",[1046],"model_aggregator",{"slug":1048,"name":1049,"description":1050,"icon_url":1051,"addition_types":1052},"build_deploy","Build and deploy","Testing, packaging, and shipping your code on every change.","https:\u002F\u002Fassets.tekyous.dev\u002Ficons\u002Faddition-groups\u002Fbuild_deploy.svg",[1053],"ci_cd",{},[1056,385],"coding_agent",{"Development Tools":1058,"LLM":1088,"AI Infrastructure":1105},[1059,1073],{"tool_id":228,"name":229,"slug":230,"tooltip_description":231,"logo_url":232,"logo_bg":60,"pricing_model":1060,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":1061,"subcategory":1062,"categories":1063,"subcategories":1065,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":1067,"aliases":1071,"score_reasonings":1072,"published_date":107,"last_updated_date":257},{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":239,"name":240,"slug":241},[1064],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[1066],{"subcategory_id":239,"name":240,"slug":241,"category_id":235,"is_primary":3,"display_order":78},[1068,1069,1070],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":252,"flexibility":253,"performance":254,"popularity":255,"portability":256},{"tool_id":504,"name":505,"slug":506,"tooltip_description":507,"logo_url":508,"logo_bg":60,"pricing_model":1074,"learning_curve_score":153,"popularity_score":43,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":1075,"subcategory":1076,"categories":1077,"subcategories":1079,"flexibility_score":38,"performance_score":153,"portability_score":46,"is_featured":81,"tags":1081,"aliases":1085,"score_reasonings":1087,"published_date":337,"last_updated_date":107},{"slug":62,"display_name":63,"description":64},{"category_id":235,"name":236,"slug":237},{"subcategory_id":512,"name":501,"slug":513},[1078],{"category_id":235,"name":236,"slug":237,"is_primary":3,"display_order":78},[1080],{"subcategory_id":512,"name":501,"slug":513,"category_id":235,"is_primary":3,"display_order":78},[1082,1083,1084],{"tag_id":520,"name":521,"slug":522,"tag_type":130},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},[1086],{"alias":527,"is_primary":3},{"learning_curve":529,"flexibility":530,"performance":531,"popularity":532,"portability":533},[1089],{"tool_id":377,"name":378,"slug":379,"tooltip_description":380,"logo_url":381,"logo_bg":115,"pricing_model":1090,"learning_curve_score":46,"popularity_score":38,"hosting_assignment_type":35,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":1091,"subcategory":1092,"categories":1093,"subcategories":1095,"flexibility_score":43,"performance_score":43,"portability_score":43,"is_featured":81,"tags":1097,"aliases":1101,"score_reasonings":1104,"published_date":337,"last_updated_date":107},{"slug":62,"display_name":63,"description":64},{"category_id":384,"name":374,"slug":385},{"subcategory_id":387,"name":388,"slug":389},[1094],{"category_id":384,"name":374,"slug":385,"is_primary":3,"display_order":78},[1096],{"subcategory_id":387,"name":388,"slug":389,"category_id":384,"is_primary":3,"display_order":78},[1098,1099,1100],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":323,"name":324,"slug":325,"tag_type":135},{"tag_id":89,"name":90,"slug":91,"tag_type":87},[1102,1103],{"alias":400,"is_primary":3},{"alias":402,"is_primary":3},{"learning_curve":404,"flexibility":405,"performance":406,"popularity":407,"portability":408},[1106],{"tool_id":55,"name":56,"slug":57,"tooltip_description":58,"logo_url":59,"logo_bg":60,"pricing_model":1107,"learning_curve_score":65,"popularity_score":43,"hosting_assignment_type":66,"hosting_provider_restriction":67,"hosting_target_restriction":67,"hosting_compatible_tool_ids":35,"parent_tool_id":35,"category":1108,"subcategory":1109,"categories":1110,"subcategories":1112,"flexibility_score":43,"performance_score":38,"portability_score":43,"is_featured":81,"tags":1114,"aliases":1119,"score_reasonings":1120,"published_date":107,"last_updated_date":108},{"slug":62,"display_name":63,"description":64},{"category_id":69,"name":70,"slug":71},{"subcategory_id":73,"name":74,"slug":75},[1111],{"category_id":69,"name":70,"slug":71,"is_primary":3,"display_order":78},[1113],{"subcategory_id":73,"name":74,"slug":75,"category_id":69,"is_primary":3,"display_order":78},[1115,1116,1117,1118],{"tag_id":84,"name":85,"slug":86,"tag_type":87},{"tag_id":89,"name":90,"slug":91,"tag_type":87},{"tag_id":93,"name":94,"slug":95,"tag_type":87},{"tag_id":97,"name":98,"slug":99,"tag_type":87},[],{"learning_curve":102,"flexibility":103,"performance":104,"popularity":105,"portability":106},[1122,1134,1144,1157,1169,1183,1199,1214,1225,1238,1252,1266],{"stack_id":1123,"slug":1124,"name":1125,"tagline":1126,"experience_level":1127,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":46,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1130},235,"qwen-code-dev-stack","Qwen Code Dev Stack","Alibaba's open-source coding agent in the terminal, desktop, or browser, running Qwen, GLM, Kimi, or any model you choose.","intermediate","developer","https:\u002F\u002Fassets.tekyous.dev\u002Ficons\u002Fstack-types\u002Fdeveloper.svg",[1131,1132,1133],{"tool_id":283,"slug":285,"name":284,"logo_url":287,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":377,"slug":379,"name":378,"logo_url":381,"logo_bg":115},{"stack_id":1135,"slug":1136,"name":1137,"tagline":1138,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":38,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1140},232,"opencode-dev-stack","OpenCode Dev Stack","OpenCode in the terminal, desktop app, or browser, running open-weight models like DeepSeek, GLM, and Kimi.","advanced",[1141,1142,1143],{"tool_id":228,"slug":230,"name":229,"logo_url":232,"logo_bg":60},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":468,"slug":470,"name":469,"logo_url":472,"logo_bg":115},{"stack_id":1145,"slug":1146,"name":1147,"tagline":1148,"experience_level":1127,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":153,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1149},233,"cline-dev-stack","Cline Dev Stack","Cline's open-source agent in the terminal or its desktop app, running GLM, Kimi, Claude, or any model you choose.",[1150,1151,1152],{"tool_id":259,"slug":261,"name":260,"logo_url":263,"logo_bg":60},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1153,"slug":1154,"name":1155,"logo_url":1156,"logo_bg":115},157,"glm","GLM","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fglm.png",{"stack_id":1158,"slug":1159,"name":1160,"tagline":1161,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":153,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1162},245,"deepseek-harness-dev-stack","DeepSeek Harness Dev Stack","DeepSeek's open-source, plugin-built agent in a local web interface or desktop app, running DeepSeek, Kimi, GLM, Claude, or GPT.",[1163,1167,1168],{"tool_id":1164,"slug":1165,"name":1166,"logo_url":472,"logo_bg":115},208,"deepseek-harness","DeepSeek Harness",{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":468,"slug":470,"name":469,"logo_url":472,"logo_bg":115},{"stack_id":1170,"slug":1171,"name":1172,"tagline":1173,"experience_level":1174,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":65,"score_learning_curve":46,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1175},234,"zcode-dev-stack","ZCode Dev Stack","Z.ai's desktop coding agent with Goal Mode, running GLM, Kimi, Claude, or another provider's model.","beginner",[1176,1181,1182],{"tool_id":1177,"slug":1178,"name":1179,"logo_url":1180,"logo_bg":115},206,"zcode","ZCode","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fzcode.png",{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1153,"slug":1154,"name":1155,"logo_url":1156,"logo_bg":115},{"stack_id":1184,"slug":1185,"name":1186,"tagline":1187,"experience_level":1174,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":43,"score_learning_curve":65,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1188},160,"cursor-dev-stack","Cursor Dev Stack","Cursor as editor and coding agent in one, with Grok by default and Claude, GPT, or Gemini a click away.",[1189,1194,1195],{"tool_id":1190,"slug":1191,"name":1192,"logo_url":1193,"logo_bg":60},18,"cursor","Cursor","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fcursor.svg",{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1184,"slug":1196,"name":1197,"logo_url":1198,"logo_bg":60},"grok","Grok","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgrok.svg",{"stack_id":598,"slug":1200,"name":1201,"tagline":1202,"experience_level":1174,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":43,"score_learning_curve":65,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1203},"vs-code-ai-dev-stack","VS Code AI Dev Stack","VS Code stays your editor; the AI extension and version control underneath are both your choice.",[1204,1208,1213],{"tool_id":384,"slug":1205,"name":1206,"logo_url":1207,"logo_bg":115},"vs-code","Visual Studio Code","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fvs-code.svg",{"tool_id":1209,"slug":1210,"name":1211,"logo_url":1212,"logo_bg":60},194,"github-copilot","GitHub Copilot","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fgithub-copilot.svg",{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"stack_id":996,"slug":1215,"name":1216,"tagline":1217,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":38,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1218},"claude-code-dev-stack","Claude Code Dev Stack","Claude Code in the terminal, desktop app, or web, paired with Claude and the add-ons to run it anywhere.",[1219,1220,1221],{"tool_id":97,"slug":340,"name":339,"logo_url":342,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1222,"slug":1223,"name":1224,"logo_url":342,"logo_bg":115},153,"claude","Claude",{"stack_id":1226,"slug":1227,"name":1228,"tagline":1229,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":38,"score_learning_curve":46,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1230},168,"codex-dev-stack","Codex Dev Stack","OpenAI Codex on your desktop, in your terminal, or in the cloud, steerable from your phone.",[1231,1232,1233],{"tool_id":308,"slug":310,"name":309,"logo_url":312,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1234,"slug":1235,"name":1236,"logo_url":1237,"logo_bg":60},155,"openai","OpenAI","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fopenai.svg",{"stack_id":1239,"slug":1240,"name":1241,"tagline":1242,"experience_level":1127,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":153,"score_learning_curve":46,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1243},229,"conductor-dev-stack","Conductor Dev Stack","Run Claude Code, Codex, and Cursor's agent in parallel on your Mac with Conductor, each in its own workspace.",[1244,1249,1250,1251],{"tool_id":1245,"slug":1246,"name":1247,"logo_url":1248,"logo_bg":60},277,"conductor","Conductor","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fconductor.png",{"tool_id":97,"slug":340,"name":339,"logo_url":342,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1222,"slug":1223,"name":1224,"logo_url":342,"logo_bg":115},{"stack_id":1253,"slug":1254,"name":1255,"tagline":1256,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":153,"score_learning_curve":153,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1257},228,"herdr-dev-stack","Herdr Dev Stack","Keep Claude Code, Codex, and other coding agents running in Herdr sessions you can reattach from any machine.",[1258,1263,1264,1265],{"tool_id":1259,"slug":1260,"name":1261,"logo_url":1262,"logo_bg":60},252,"herdr","Herdr","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Fherdr.svg",{"tool_id":97,"slug":340,"name":339,"logo_url":342,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1222,"slug":1223,"name":1224,"logo_url":342,"logo_bg":115},{"stack_id":1267,"slug":1268,"name":1269,"tagline":1270,"experience_level":1139,"project_type":35,"stack_type_slug":1128,"stack_type_icon_url":1129,"score_popularity":153,"score_learning_curve":38,"catalog_display_order":35,"published_date":35,"last_updated_date":35,"core_tool_previews":1271},227,"orca-dev-stack","Orca Dev Stack","Run Claude Code, Codex, and other coding agents side by side in Orca, each on its own branch and model.",[1272,1277,1278,1279],{"tool_id":1273,"slug":1274,"name":1275,"logo_url":1276,"logo_bg":60},255,"orca","Orca","https:\u002F\u002Fassets.tekyous.dev\u002Flogos\u002Ftools\u002Forca.png",{"tool_id":97,"slug":340,"name":339,"logo_url":342,"logo_bg":115},{"tool_id":504,"slug":506,"name":505,"logo_url":508,"logo_bg":60},{"tool_id":1222,"slug":1223,"name":1224,"logo_url":342,"logo_bg":115},[1281,1284,1287,1290,1293,1296],{"question":1282,"answer":1283},"Can the agent keep working when my laptop is closed?","Only if the machine doing the work stays on. There is no vendor cloud in this stack: the model and the agent both run on hardware you control. The usual answer is two machines. A desktop with the graphics card runs the model and stays on, and the laptop runs the agent and reaches the model over the home network, or over Tailscale from anywhere. With the laptop closed, the model keeps serving, but the agent session on the laptop stops. For work that continues while you are away, the agent has to run on the always-on machine as well, and OpenCode's web interface then shows its sessions in a phone browser. A rented GPU server does the same job without hardware at home, at a much higher monthly cost.",{"question":1285,"answer":1286},"Which agents work with which runtime?","All five agents work with Ollama, and most work with the others. OpenCode documents Ollama, LM Studio, and llama.cpp as providers. Cline has Ollama and LM Studio built in. Qwen Code takes any runtime that offers an OpenAI-compatible address, which all four do. Codex's open-model mode targets Ollama and LM Studio. Claude Code needs an Anthropic-compatible endpoint, which Ollama, LM Studio, and llama.cpp provide. Ollama and Unsloth each have a single command that starts OpenCode, Claude Code, or Codex already connected, which is also how those three reach Unsloth. Two cautions apply to the vendor agents: Anthropic supports Claude Code with Claude models only, and Codex is tuned for [OpenAI](\u002Ftools\u002Fopenai)'s models, so both lose some features and reliability on a local model. The models are a separate choice and every runtime here can load the same open weights.",{"question":1288,"answer":1289},"Is a local model good enough to replace a coding subscription?","For some work, not for all of it. A coding model that fits one graphics card handles contained tasks well: writing a function, adding tests, explaining code, a refactor inside a few files. It falls behind hosted frontier models on long tasks that span many files, where it loses track, repeats a failed edit, or calls a tool wrongly. It is also slower. Most people who run this stack keep both: the local model for routine and private work, and a hosted model for the hard problems, either through a subscription agent or through a gateway from the same agent. If privacy is the reason, the question is different, and a weaker model that keeps the code in house is the right trade.",{"question":1291,"answer":1292},"What hardware do I need, and what does it cost?","The coding models here are 15 to 19 GB downloads, and a model needs about its file size in graphics memory plus room for the context window. In practice that means a graphics card with 24 GB, or an Apple Silicon Mac with 32 GB or more of unified memory. Cline's docs give a similar ladder for system memory: 16 to 32 GB for small or quantized models, 32 to 64 GB for mid-size coding models, and more for large ones. A card with 12 or 16 GB runs smaller models for simpler tasks. The software is all free, so the cost is the hardware and the electricity. Renting is possible, from about €214 a month for a dedicated GPU server. The Pricing section has the details.",{"question":1294,"answer":1295},"How large a context window do I need?","Larger than the runtime gives you by default, and this is the setting that decides whether the stack works at all. A coding agent sends its instructions, its tool definitions, and the files it is reading on every turn. Ollama's docs recommend at least 64,000 tokens for coding tools, while its default on a card with less than 24 GB is about 4,000, at which point the agent's own instructions no longer fit and tool calls fail. LM Studio's docs ask for more than about 25,000 for Claude Code and Codex. Every extra token of context uses memory on top of the model itself, so the real choice is between a larger model with a tight context and a smaller one with room to read. For agent work, the room usually matters more.",{"question":1297,"answer":1298},"Can I run several coding agents at the same time on one local model?","Yes, in two ways. The simple one is separate sessions that you manage yourself: OpenCode in one terminal, Cline in the editor, Claude Code in another terminal, each pointed at the same local address. The other is an agentic dev environment, and how it connects depends on the kind. [Herdr](\u002Ftools\u002Fherdr) and [Orca](\u002Ftools\u002Forca) run every agent in a real terminal, so whatever starts an agent on the local model in your own terminal, a runtime's launch command or the agent's own settings, works the same there, and Orca can store that command per agent. [T3 Code](\u002Ftools\u002Ft3-code) and [Paseo](\u002Ftools\u002Fpaseo) start the agents themselves, so the local address goes into their provider settings as a custom endpoint, a setup both document. The limit is the hardware, not the software. By default Ollama answers one request per model at a time and queues the rest, and each extra parallel slot multiplies the memory the context needs, so on one graphics card the agents mostly take turns.",{"summary":1300,"starting_cost_label":1301,"has_free_tier":3,"line_items":1302},"Every tool in this stack is free: the runtimes, the agents, and the open-weight models. The cost is hardware. A graphics card with 24 GB of memory, or a Mac with 32 GB or more, runs the coding models, and electricity is the only running cost. Renting a GPU server starts at about €214 a month, or under a dollar an hour for a cloud GPU you switch off between sessions. A hosted model through a gateway is optional and billed per token.","Free on hardware you own",[1303,1307,1311,1315,1319,1323,1327],{"label":1304,"cost":1305,"note":1306},"Runtime and agent","Free (open source)","Ollama, llama.cpp, and Unsloth are open source and LM Studio is free to use. OpenCode, Cline, Qwen Code, and the Codex CLI are open source.",{"label":1308,"cost":1309,"note":1310},"Models","Free (open weights)","Downloaded once, 9 to 19 GB each for the coding models. Licences vary by family, mostly Apache 2.0 or MIT.",{"label":1312,"cost":1313,"note":1314},"Hardware (your own computer)","Owned, plus electricity","About 24 GB of graphics memory, or a Mac with 32 GB or more, for the 15 to 19 GB coding models.",{"label":1316,"cost":1317,"note":1318},"Version control (GitHub or GitLab)","Free–$29\u002Fmo","The free tiers cover individual developers; GitHub Team is $4 a month and GitLab Premium $29 per seat.",{"label":1320,"cost":1321,"note":1322},"Hosted fallback (optional)","Provider prices + 0–5.5%","OpenRouter charges provider prices plus 5.5% on credit purchases; LiteLLM is free to self-host.",{"label":1324,"cost":1325,"note":1326},"Server (optional)","From ~€214\u002Fmo, or ~$0.76\u002Fhour","A dedicated server with a 24 GB card and 64 GB of RAM at a fixed monthly price, or a cloud GPU with 20 GB billed by the hour.",{"label":1328,"cost":1329,"note":1330},"Remote access (optional)","Free","Tailscale's Personal plan is free.","2026-10-16",{"title":1333,"description":1334,"og_image":35,"canonical":1335},"Local LLM Dev Stack: Tools, Pricing & How to Deploy | Tekyous","A coding agent on a model that runs on your own machine, with no code sent to a provider. Compare Local LLM Dev Stack tools, pricing & how to deploy on Tekyous.","https:\u002F\u002Ftekyous.dev\u002Fstacks\u002Flocal-llm-dev-stack",[],1792744070359]