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
- [Bug]: ModelResponse dropped from history when ModelRetry crosses UsageLimits
- `AnthropicModel` on `AsyncAnthropicBedrock` prices `global.` inference-profile requests at the geographic rate, because the price lookup uses the response's bare model name
- Let `ClearToolResults` keep the newest N results per named tool (`keep_per_tool`)
- Make the v2.60 `RetryPromptPart` change findable and detectable for code that only reads message parts
- Harness `Skills`: point packaged skills without a workspace to deferred capabilities, and show a database-backed runbook
- `AnthropicModel.count_tokens` on legacy `AsyncAnthropicBedrock` fails for deferred tools, tool-search history and inference-profile ids
- `OpenAIJsonSchemaTransformer` adds `additionalProperties: false` to free-form objects in non-strict tools
- MCPToolset unwraps a genuine `result` field from structured tool output
- Send AG-UI `RunAgentInput.state` to the model as `StructuredContent` in the user turn
- Cap fallback delays in `wait_retry_after` at `max_wait`
- Docs
- Python not yet supported