pydantic-ai
https://github.com/pydantic/pydantic-ai
Python
GenAI Agent Framework, the Pydantic way
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported6 Subscribers
Add a CodeTriage badge to pydantic-ai
Help out
- Issues
- End a run with a final tool-less answer instead of raising when a usage limit is reached
- Give custom streaming models a public alternative to `ModelResponsePartsManager`
- Harness `SpendLimits`: let a `Budget` ceiling be resolved per scope
- Add an Amazon Bedrock AgentCore Code Interpreter sandbox capability (`AgentCoreSandbox`)
- `web_fetch` returns `BinaryContent` with an invalid or unsendable media type, and `OpenAIChatModel` rejects it with `RuntimeError`
- Code mode: run_code keeps using the session after Monty kills the worker for memory
- Fix `TestModel` streamed tool-call usage reporting zero output tokens
- Let users override or fail open on capability gating for model ids our profiles don't know yet
- Warn when `thinking` is dropped because the resolved profile can't honour it
- Give tool and output retry exhaustion their own `UnexpectedModelBehavior` subclasses
- Docs
- Python not yet supported