Getting Started
Welcome! This section gets you from zero to a running agent as fast as possible.
Before You Start
- Python 3.10+ (up to 3.14)
- An LLM API endpoint — OpenAI-compatible, Anthropic, or any supported adapter
- A few MB of RAM for the framework itself (context memory grows with your session)
No prior knowledge of AmritaSense is required for this section — everything you need is covered inline.
Install
We recommend uv for environment management:
bash
uv init
uv venv
uv add amrita-coreOr with pip:
bash
pip install amrita-coreWorking directly from the source repository? Clone it and pip install -e . (see the repo README.md).
What's Next
- Minimal Example — a 10-line runnable agent
- Basic Example — streaming, tools and sessions in one place
- Then follow the Tutorials to build up systematically
