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
- Harness Shell: a repeated or empty tool_call_id in one run makes the second command silently return the first one's output
- AG-UI `dump_messages`/`load_messages` round trip drops Gemini thought signatures (`provider_details`), so replayed history sends `skip_thought_signature_validator`
- Bedrock Mantle tool-call IDs exceed Azure’s 64-character limit during cross-provider replay
- `SummarizingCompaction(keep_user_messages=True)` drops tail messages from both the summary and the kept history
- TestModel: streamed tool calls report 0 output tokens, unlike run()
- Stop IDEs auto-importing `RunContext` from `pydantic_ai._run_context`, and decide on private APIs downstream depends on
- Document MCP stdio server stderr and `init_timeout` for slow cold starts
- Harness Shell: continuing a conversation after a crash runs the interrupted command again
- Report a harness `Shell` job whose processes died as lost, and don't signal a process that reused its PID
- cache-write span attribute doesn't match the current OTel GenAI convention
- Docs
- Python not yet supported