docsfy

Overview of docsfy, what it does, how it uses AI CLI tools to generate Mintlify-quality documentation from GitHub repositories, and the core value proposition

Step-by-step guide to get docsfy running locally with Docker Compose, generate documentation for your first repository, and view the results

Detailed installation options including Docker-based deployment, local development setup with uv and Python 3.12+, and system prerequisites like git, Node.js, and AI CLI tools

Complete reference of all environment variables including AI_PROVIDER, AI_MODEL, AI_CLI_TIMEOUT, authentication keys (ANTHROPIC_API_KEY, GEMINI_API_KEY, CURSOR_API_KEY), Vertex AI settings, and LOG_LEVEL

How to configure and switch between supported AI providers: Claude Code, Gemini CLI, and Cursor Agent, including binary requirements, authentication methods, and model selection

How docsfy stores data using SQLite for metadata and the filesystem for generated documentation, including the directory structure under /data/projects/ and database schema

How to submit a repository for documentation generation via the POST /api/generate endpoint, monitor generation status, and understand the four-stage pipeline (clone, plan, content, render)

How to access generated documentation through the built-in static file server at /docs/{project}/, including navigation, dark/light theme toggle, and client-side search

How to download generated documentation as a .tar.gz archive via the API and host it on your own infrastructure, CDN, or static hosting service

How docsfy detects repository changes via commit SHA tracking, re-evaluates documentation plans, and regenerates only affected pages for efficient updates

Deep dive into the four-stage generation pipeline: repository cloning, AI-driven documentation planning (plan.json), concurrent page content generation with semaphore-limited concurrency, and HTML rendering with Jinja2

How the HTML renderer converts markdown and plan.json into static sites with sidebar navigation, syntax highlighting via highlight.js, callout boxes, card layouts, responsive design, and search indexing

Technical details of the provider pattern for AI CLI tools, subprocess invocation, async execution via asyncio.to_thread, multi-strategy JSON response parsing, and availability checking

Start documentation generation for a GitHub repository URL, including request format, response structure, and supported repository types (public/private, SSH/HTTPS)

List all projects and their generation status, get detailed project information including last generated timestamp, commit SHA, and page list, and delete projects

Download a project's generated documentation site as a .tar.gz archive for self-hosting

Health check endpoint for monitoring and container orchestration

Production deployment using the multi-stage Dockerfile with python:3.12-slim, non-root user configuration for OpenShift compatibility, volume mounts for data persistence and cloud credentials, and health check configuration

Running docsfy with docker-compose.yaml including port mapping, environment file configuration, volume mounts for data and gcloud credentials, and health check settings

Setting up a local development environment with uv, Python 3.12+, pre-commit hooks (ruff, mypy, flake8, gitleaks, detect-secrets), and tox for testing

Guidelines for contributing to docsfy including code style enforcement via pre-commit, running tests, build system (hatchling), and project conventions