AI Docs Assistant

An intelligent API documentation assistant that combines RAG (Retrieval-Augmented Generation), vector search, and LLM agents to automate API documentation workflows. The system can search existing documentation, generate new documentation on demand, and automatically validate the output.

✨ Features

• 🔍 Semantic search with Qdrant vector similarity
• 🤖 AI generation using CrewAI agents with Ollama LLMs
• ✅ Auto-validation to ensure consistent documentation format
• 📝 Smart storage with automatic document naming and organization
• 🏥 Health checks for Qdrant, Ollama, and the RAG pipeline
• 🔄 Incremental indexing for adding new docs without a full rebuild
• 🐳 Docker ready for easy deployment

🏗 Architecture

User Query → /search → Found? → Return existing doc
                    ↓ No
              /generate → Generator Agent → Validator Agent
                                            ↓
                                      Save to docs/
                                            ↓
                                    Index in Qdrant

🛠 Tech Stack

| Component | Technology | |---------------------|----------------------------| | API Framework | FastAPI | | Vector Database | Qdrant | | LLM Runner | Ollama | | Embeddings | mxbai-embed-large | | LLM Model | mistral:latest | | Multi-Agent | CrewAI | | Document Processing | LangChain | | Container | Docker + docker-compose |

📋 Prerequisites

• Python 3.11+
• Docker & Docker Compose
• Ollama installed locally or accessible via network
• Required Ollama models pulled:

ollama pull mistral:latest
ollama pull mxbai-embed-large

🚀 Quick Start

Option 1: Docker Compose (Recommended)

git clone https://github.com/yourusername/ai-docs-assistant.git
cd ai-docs-assistant
cp .env.example .env
# Edit .env with your settings
docker-compose up -d

• Place your initial .md documentation files in the docs/ directory.
• The system will index documents automatically on startup.
• Access the API at http://localhost:8000.
• Health check: http://localhost:8000/health

Option 2: Local Development

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt
docker run -p 6333:6333 qdrant/qdrant
cp .env.example .env
# Update OLLAMA_HOST=localhost if running locally
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

📌 API Endpoints

GET /health

Returns a health summary for all services.

{
  "status": "healthy",
  "checks": {
    "qdrant": true,
    "ollama": true,
    "docs": true,
    "rag_canary": true
  }
}

POST /search

Perform semantic search for documentation.

Request

{
  "query": "Get user profile endpoint"
}

Response (found)

{
  "found": true,
  "content": "### GET /api/v1/profile\n**Description**: ...",
  "message": null
}

Response (not found)

{
  "found": false,
  "content": null,
  "message": "Documentation not found. Use /generate to create new documentation."
}

POST /generate

Generate new API documentation automatically.

Request

{
  "query": "Create new task for user"
}

Response (success)

{
  "success": true,
  "message": "Document successfully created and saved",
  "content": "### POST /api/v1/tasks\n**Description**: ...",
  "file_path": "docs/create_task.md"
}

📂 Project Structure

ai-docs-assistant/
├── app/
│   ├── __init__.py
│   ├── agents.py          # CrewAI agents (generator + validator)
│   ├── health.py          # Health check endpoints
│   ├── logger.py          # Logging configuration
│   ├── main.py            # FastAPI application
│   ├── rag.py             # Qdrant + LangChain integration
│   ├── schemas.py         # Pydantic models
│   ├── settings.py       # Environment configuration
│   └── storage.py        # File storage utilities
├── docs/                  # Documentation storage (auto-created)
├── logs/                  # Application logs (auto-created)
├── tests/
│   └── test_endpoints.py  # Pytest test suite
├── .env.example           # Environment variables template
├── docker-compose.yml     # Docker orchestration
├── Dockerfile             # Container definition
├── requirements.txt       # Python dependencies
└── README.md              # This file

⚙️ Configuration

| Variable | Description | Default | |------------------------|-------------------------------------|-------------------------------| | QDRANT_HOST | Qdrant server host | qdrant | | QDRANT_PORT | Qdrant server port | 6333 | | QDRANT_COLLECTION_NAME | Vector collection name | api_docs | | EMBEDDING_MODEL_NAME | Ollama embedding model | mxbai-embed-large | | VECTOR_SIZE | Embedding dimension | 1024 | | API_KEY | Ollama API key (if required) | ollama | | OLLAMA_HOST | Ollama server host | host.docker.internal | | OLLAMA_PORT | Ollama server port | 11434 | | OLLAMA_MODEL | LLM model for generation | ollama/mistral:latest |

📝 Documentation Format

The generator agent produces documentation in this strict format:

### METHOD /endpoint
**Description**: Brief description of the endpoint
**Parameters**: List of required/optional parameters
**Response**:
{
  "example": "response"
}

🧪 Testing

Install pytest and run the test suite:

pip install pytest
pytest tests/test_endpoints.py -v

Tests include:

• Semantic search for existing documents
• Search for non-existent documents
• New document generation
• Health check endpoint

📈 Troubleshooting

Qdrant connection issues

docker ps | grep qdrant
docker logs ai-docs-assistant-qdrant-1

Ollama connection issues

curl http://localhost:11434/api/tags
ollama pull mistral:latest
ollama pull mxbai-embed-large

Empty search results

• Ensure docs/ contains .md files
• Check logs for indexing errors
• Verify embedding model matches VECTOR_SIZE (1024 for mxbai-embed-large)

🌱 Future Improvements

• Support for multiple document formats (PDF, OpenAPI, JSON)
• Document versioning
• Authentication / authorization
• Web UI for documentation browsing
• Export to Swagger / OpenAPI format
• Batch document processing
• Custom prompt templates per API style

📄 License

MIT

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch: git checkout -b feature/amazing-feature
3. Commit your changes: git commit -m 'Add amazing feature'
4. Push to the branch: git push origin feature/amazing-feature
5. Open a Pull Request