AGI Agentic 🚀

Documentation | Quick start | Join Discord | Blog | NewsLetter | Careers

Demo

Video

Images

Key Features

• 🎯 ReAct
• 🧪 MCP
• 🔗 Tools
• 📊 Agent

🔥 Quickstart (uv-based)

This Quickstart uses the project's uv workflow (preferred). Commands assume macOS and a POSIX shell.

Prerequisites

• Python 3.11+ (3.12 recommended)
• Git
• uv CLI (project package manager). If you don't have it, install per your environment (e.g. pip install uv).

1. Clone repository

git clone https://github.com/lifangjunone/AGIAgentic.git
cd AGIAgentic/AGIAgentic-backend

2a) (Optional) Create and activate a venv (recommended if you prefer)

python -m venv .venv
source .venv/bin/activate

2b) (If you rely on uv-managed env, skip venv creation)

3. Sync dependencies with uv (reads uv.lock / pyproject)

# synchronize / install pinned deps from uv.lock
uv sync

4. Configure environment variables

# environment config
export RUN_ENV=development|production|test
# config file path:
├─ envs/
│  ├─ development.env # dev env
│  ├─ production.env # prod env
│  ├─ test.env # test env

Or export for current shell:

export OPENAI_API_KEY="sk-..."

5. Start the backend (preferred: uv)

# run the app via project CLI (hot-reload behaviour depends on project)
uv run main.py

Fallback: uvicorn (if uv is unavailable)

python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000

6. Test the streaming plan API (example)

curl -X 'POST' \
  'http://127.0.0.1:8000/api/v1/plan_executor/stream' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "user_task": "昨天北京到天津的车票都有哪些时间的",
  "user_id": "001"
}'

The endpoint returns Server-Sent Events (SSE). Parse each "data:" line as JSON.

7. Run tests

pytest -q

Troubleshooting (common)

• "Expected dict, got <...>": ensure runtime events/state only contain serializable types (strings/dicts/lists/primitives). Avoid writing langchain HumanMessage / complex objects into graph state or SSE payloads.
• Dependency / lock issues: run uv sync again, or recreate your venv and re-run uv sync.
• Model/tool errors: verify API keys and tool configurations in .env.

If you want, I can add platform-specific launch scripts, a systemd service example, or a docker-compose file next.

We welcome contributions from the community! Whether it's bug fixes, feature additions, or documentation improvements, your input is valuable.

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

Cite Us

@misc{ragas2024,
  author       = {LI},
  title        = {AGIAgentic: Plan and execute},
  year         = {2025},
  howpublished = {\url{https://github.com}},
}

Technical documentation

+Jump to the streaming events reference: astream_events — API Reference