Software Project Types for Tech Stack Building

Project types are used in the Tekyous stack builder's Questionnaire mode to tailor tool recommendations to the kind of project you are building. When you select a project type, the recommendation engine applies a set of curated rules to surface the most relevant tools for each category in your stack.

Project types apply to project tech stacks specifically — they describe what is being built, not how the developer works or what infrastructure runs underneath. The stack builder also supports developer stacks and infrastructure stacks, but those are not organised by project type.

It is also important to note that project types are not mutually exclusive — they are often complementary. An API is commonly a part of a web application, and a dashboard will typically consume data flowing through an ETL pipeline. Even though the builder treats these as separate project types, the final tech stack might be a combination of several of them, and may also include tools from developer and infrastructure stacks.

There are currently 9 project types. Click any project type below to open the stack builder in Questionnaire mode with that type pre-selected.

---

Web Application

A browser-based web app or SaaS product. This is the broadest project type and the most commonly used starting point. Web applications can be built across a wide spectrum of approaches:

• Code-based: using frontend frameworks, backend languages and server frameworks, and a database of your choice — giving you full control over every layer of the architecture
• No-code / low-code: using visual app builders and managed platforms that handle logic, data, and UI without writing code from scratch
• Hybrid: combining a managed backend-as-a-service (BaaS) for auth, storage, and data with a fully custom frontend — reducing backend complexity without sacrificing frontend control

---

Website / Landing

A marketing website, blog, portfolio, or documentation site. The key distinction from a web application is that websites are primarily content-driven — they typically do not require user accounts, real-time features, or complex backend logic. Most likely will not need a dedicated backend service and a separate database, but might rely on a CMS instead. The typical tools in this space are static site generators, CMS platforms, and headless content tools — paired with lightweight hosting suited to content-first publishing. No-code and low-code website builders are also a strong fit here, particularly for teams that want fast time-to-publish without involving developers.

---

Mobile App

An iOS or Android app, built either natively or with a cross-platform framework. Native development gives maximum platform control, while cross-platform frameworks let you target both stores from a single codebase at the cost of some platform-specific depth. Mobile apps typically need a backend to handle user data, authentication, and real-time updates — making backend-as-a-service a core part of the mobile stack alongside the framework itself. Mobile-specific concerns like push notifications, offline support, biometric authentication, and app store distribution shape the tool landscape in ways that differ significantly from web or API projects.

---

API / Backend

A REST or GraphQL API with no frontend. This project type often serves as the backbone for web apps, mobile apps, or third-party integrations — built and managed independently from any consumer-facing interface. The stack is server-side only — server frameworks, databases, authentication, and infrastructure — with no frontend tooling involved. Deployment decisions (serverless functions, containerised services, or traditional servers) significantly shape the relevant tools, as each model comes with different runtime, scaling, and operational tradeoffs. Rate limiting, API documentation, and access management are also considerations specific to this project type.

---

Automation / Bot

Scripts, workflows, bots, and scheduled jobs. Covers deterministic, rule-based automations — connecting systems, triggering on events, and processing data in sequence. This project type spans a particularly wide range of implementation approaches: no-code platforms let non-technical users build integrations visually, while code-based tools like cron jobs, webhook handlers, or custom scripts give developers precise control. Common use cases include syncing data between services, sending notifications, automating repetitive admin tasks, and operating bots on chat platforms or social channels.

Note: AI agent workflows will be covered by a dedicated project type when introduced.

---

Data Pipeline / ETL

Building data ingestion, transformation, or orchestration workflows. A data pipeline moves data from one or more source systems, applies transformations to clean, aggregate, or enrich it, and loads the result into a destination — typically a data warehouse or analytical store. Pipelines can run on a schedule, trigger on events, or stream data in near-real time depending on the use case. This project type frequently feeds downstream dashboards or ML training workflows, making it a common component in larger data-oriented architectures.

---

Dashboard / Reporting

Analytics dashboards, BI reports, or embedded data visualisations. Dashboards sit at the end of a data chain — they consume structured data from databases, warehouses, or APIs and present it in charts, tables, and summary views. The right tool depends heavily on the audience: dedicated BI platforms work well for business users who need self-serve reporting, while code-based data app frameworks or notebook environments give analysts and developers more flexibility and customisation. Embedded analytics — integrating charts directly into a product UI — requires a different approach again, typically closer to a frontend component library than a BI platform.

---

ML / AI Model

Training, evaluating, or deploying machine learning and AI models. Covers classical ML and deep learning pipelines — data processing, feature engineering, model training, experiment tracking, and model serving. The tooling landscape differs significantly depending on whether you are working with classical ML (structured data, tabular models) or deep learning (neural networks, large datasets, GPU compute). Experiment tracking and reproducibility are first-class concerns, and model serving introduces its own infrastructure decisions separate from training. Covers the full model lifecycle from raw data to a deployed inference endpoint.

Note: AI agent project types will be introduced separately, with coverage for both no-code and code-first approaches.

---

Browser Extension

A browser extension or add-on targeting Chrome, Firefox, or other browsers. Extensions run inside the browser and can read and modify page content, intercept network requests, inject custom UI, and access browser storage — capabilities unavailable to regular web pages. The tooling follows the Manifest V3 architecture, which separates content scripts (running in the page context) from background service workers (running persistently in the browser background). While extensions are often a standalone product, they frequently complement a web app or API — for example, a SaaS product with a companion extension that enhances the in-browser experience. Cross-browser compatibility is a practical concern, as Chrome and Firefox implement extension APIs with subtle differences.

---

If you are not sure which type fits your project, open the Questionnaire mode directly — the descriptions are shown inline as you select. For a broader overview of how project types work inside the platform, see the How Tekyous Works guide.