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 supported2 Subscribers
Add a CodeTriage badge to pydantic-ai
Help out
- Issues
- feat(usage): add `per_request_input_tokens_limit` to `UsageLimits`
- Unify AG-UI interrupt resume payload via a Pydantic model (`_ResumePayload`)
- Durable-exec MCP `get_instructions` has the same replay-nondeterminism bug class as #5875
- DBOS id-less MCP leaf toolsets collide on the per-run tool-defs cache key
- Cerebras: wire `<think>`-tag reasoning replay for `qwen` models (follow-up to #5891)
- UsageLimits: add a per-request input-token (context size) limit
- Type `TemporalProviderFactory` against the agent deps type instead of hardcoded `RunContext[None]`
- Fix `SpanQuery` `has_attributes` matching for JSON-serialized dict and list values
- [roundtrip-sweep] AG-UI adapter: ToolReturnPart.outcome silently dropped on round-trip (failed/denied load as success)
- Anthropic mapper re-renders unsigned/foreign `ThinkingPart`s as literal `<thinking>` text in the prompt — model mimics it and leaks reasoning into user-visible answers
- Docs
- Python not yet supported