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
- docs: document JZSTOKEN as an OpenAI-compatible model provider
- clai REPL crashes on unknown slash commands or error text that looks like Rich markup
- Raise `ModelContextWindowExceededError` for Mistral's 429 code `1300` when a single request exceeds the token budget
- Raise `ModelQuotaExceededError` for Anthropic spend-limit errors instead of a rate-limit or rejected-request error
- `max_tokens` is silently ignored on Ollama because `OllamaProvider` sends `max_completion_tokens`
- `deepseek_model_profile` misses DeepSeek V4.1 ids, so `thinking` is dropped and Together's forced-tool-choice guard is skipped
- Normalize model-id version spellings (`v4.1`, `v4p1`, `v4-1`, `5p3`) before family profile matching
- Plan: generate `OpenRouterProvider` profile flags from a committed snapshot of OpenRouter's model catalog
- `SummarizingCompaction`: trim summarizer input to fit a smaller summarizer model's context window
- A response that ends with `finish_reason='error'` and no usable output is treated as a successful response
- Docs
- Python not yet supported