Apimatic 0.1.3

Universal API documentation generator

Apimatic 🚀

A tool to automatically generate beautiful and comprehensive API documentation (Markdown/OpenAPI) from your source code.
Supports Flask, FastAPI, and more frameworks with optional AI-powered enhancements via Ollama.

---

📦 Installation

pip install apimatic

Upgrade to the latest version:

pip install --upgrade apimatic

---

⚡ Usage

Apimatic [-h] [--src SRC] [--framework [FRAMEWORK ...]] [--format {markdown,openapi}] 
         [--output OUTPUT] [--use-ollama] [--model MODEL]

---

🔑 Options

Option	Description
-h, --help	Show help message and exit
--src SRC	Root directory of the project to scan (Default: current directory)
--framework [FRAMEWORK ...]	Force a specific framework (flask, fastapi, etc.). If omitted, auto-detected
--format {markdown,openapi}	Output format (markdown or openapi) – Default: markdown
--output OUTPUT	Path for the generated output file (Default: API_Docs.md or openapi.yaml)
--use-ollama	Enhance generated docs with descriptions from a local Ollama model
--model MODEL	Ollama model for enhancement (e.g., llama3:instruct). Requires --use-ollama

---

📝 Examples

Generate Markdown docs from the current project:

Apimatic --src . --format markdown --output API_Docs.md

Generate OpenAPI spec:

Apimatic --src . --format openapi --output openapi.yaml

Force framework detection (Flask):

Apimatic --src ./my_flask_app --framework flask

Enhance documentation with AI (Ollama model):

Apimatic --src . --use-ollama --model llama3.2:1b

---

🤖 Recommended Ollama Models (1–2 GB)

When using --use-ollama, you can choose a local model for API explanations. Here are lightweight models that run well (1–2 GB range):

Model	Size	Why Use It
llama3.2:1b	~1.3GB	Fast, nimble, and great for generating clear API explanations (recommended)
gemma2:2b	~1.6GB	Slightly larger, richer outputs, good balance of quality and size
dolphin-phi	~1.6GB	Alternative small model with solid reasoning ability
orca-mini	~1.9GB	Bigger (3B params) but still under 2GB; more context-aware
moondream2	~0.8GB	Ultra-light, very fast, but less detailed

👉 Recommended Default: llama3.2:1b – best speed + clarity tradeoff.

Example:

Apimatic --src . --use-ollama --model llama3.2:1b

---

🤝 Contributing

Contributions are welcome! Please fork the repo, make your changes, and submit a PR.

---

📄 License

This project is licensed under the MIT License.