3 Releases published by 2 people

• v0.3.3 v0.3.3 (2025-06-24)

  published Jun 24, 2025

• v0.3.4 v0.3.4 (2025-06-26)

  published Jun 26, 2025

• v0.3.5 v0.3.5 (2025-06-30)

  published Jun 30, 2025

18 Pull requests merged by 11 people

• Add Slack Lead Qualifier example

  #2079 merged Jun 28, 2025

• Add ability to include snippets in docs with inline-named sections for fragments and highlighting

  #2088 merged Jun 27, 2025

• Let tools return ToolReturn to pass additional content to model, or attach metadata that's not passed to the model

  #2060 merged Jun 27, 2025

• Support strict mode in NativeOutput

  #2084 merged Jun 27, 2025

• Fix deprecated kwargs validation to prevent silent failures

  #2047 merged Jun 27, 2025

• Move ThinkingPart to preceed TextPart in OpenAIResponsesModel

  #2043 merged Jun 27, 2025

• Set 'us-central1' by default on GoogleProvider

  #2031 merged Jun 27, 2025

• fix: docs examples python version in tests

  #2069 merged Jun 27, 2025

• fix: async fixtures in conftest.py

  #2068 merged Jun 26, 2025

• Fix deprecation warning under Pydantic 2.11

  #2076 merged Jun 26, 2025

• Add progress bar on evaluate

  #1871 merged Jun 26, 2025

• docs: document how to use a custom AsyncOpenAI in the provider

  #2073 merged Jun 25, 2025

• fix(anthropic): send ThinkingPart back when using tool calls

  #2072 merged Jun 25, 2025

• Scrubbing sensitive content

  #2014 merged Jun 25, 2025

• Support NativeOutput and PromptedOutput modes in addition to ToolOutput

  #1628 merged Jun 24, 2025

• Make Edge hashable

  #2064 merged Jun 24, 2025

• note on self-hosting logfire

  #2063 merged Jun 24, 2025

• document alternative otel backends

  #2062 merged Jun 24, 2025

9 Pull requests opened by 8 people

• Fix bug with google model safety handling

  #2066 opened Jun 24, 2025

• Make it possible to stream structured while using `.iter`

  #2078 opened Jun 25, 2025

• Adding tool response in running tool spans

  #2081 opened Jun 26, 2025

• Dockerise the MCP Run Python server (#1837)

  #2090 opened Jun 27, 2025

• Implemented a convenient way to use ACI.dev Tools in PydanticAI

  #2093 opened Jun 28, 2025

• Update tests to be compatible with new OpenAI, MistralAI and MCP versions

  #2094 opened Jun 28, 2025

• Update starlette subdomain in docs

  #2099 opened Jun 30, 2025

• Bugfix: avoid race condition when refreshing google token

  #2100 opened Jun 30, 2025

• feat: AG-UI adapter (toolsets)

  #2101 opened Jun 30, 2025

20 Issues closed by 11 people

• Support Anthropic Vertex

  #2092 closed Jun 27, 2025

• How to store 'artifacts' with tool responses?

  #2034 closed Jun 27, 2025

• Support strict mode in new NativeOutput

  #2080 closed Jun 27, 2025

• Gotcha: usage_limits must be in run* function

  #1987 closed Jun 27, 2025

• google-vertex provider now uses google-genai but fails due to no default location

  #2023 closed Jun 27, 2025

• OpenAI response timestamp contradicts chronological order of message history

  #2071 closed Jun 26, 2025

• Tests errors when using pytest>=8.4.0

  #2065 closed Jun 26, 2025

• `agent.run_sync` gives a RuntimeError in a `ipynb` but works fine as a python file.

  #2077 closed Jun 25, 2025

• Unable to use Thinking AND tool calling with AnthropicModel

  #2075 closed Jun 25, 2025

• Static system prompt is not replaced by dynamic system prompt if passing message history

  #1022 closed Jun 25, 2025

• RuntimeError: anext(): asynchronous generator is already running

  #1543 closed Jun 25, 2025

• MCP improvement: client support for sampling

  #1269 closed Jun 25, 2025

• Error when using Dict[str, Any] in Pydantic models with Gemini: "should be non-empty for OBJECT type"

  #1203 closed Jun 25, 2025

• Anthropic extended thinking: invalid request error when using tools

  #2040 closed Jun 25, 2025

• Add a setting to remove prompts and completions from tracing

  #1571 closed Jun 25, 2025

• Structured outputs as an alternative to Tool Calling

  #582 closed Jun 24, 2025

• Is it possible to change result_type of agent in a tool ?

  #2061 closed Jun 24, 2025

• (openai.APITimeoutError) OpenAI requests time out, but Gemini/Anthropic requests work fine

  #774 closed Jun 24, 2025

• Can't reproduce Annotated mermaid figure

  #2046 closed Jun 24, 2025

• How to see a full LLM request body?

  #2045 closed Jun 24, 2025

13 Issues opened by 13 people

• Add support for predicted outputs in OpenAIModelSettings

  #2098 opened Jun 30, 2025

• Gemini 2.5 Pro Streamed response has no content field (just streamed thinking parts)

  #2097 opened Jun 29, 2025

• sync method for direct api `model_request_stream`

  #2096 opened Jun 29, 2025

• Save `history_processors`'s result for next round model request

  #2095 opened Jun 28, 2025

• Running many models on Bedrock with `output_type` fails with "This model doesn't support the toolConfig.toolChoice.any field."

  #2091 opened Jun 27, 2025

• BinaryContent is naively parsed when include_input is used in LLMJudge

  #2089 opened Jun 27, 2025

• Improve error message when running agent without prompt

  #2087 opened Jun 26, 2025

• Support `o3-deep-research` & `o4-mini-deep-research`

  #2086 opened Jun 26, 2025

• Feature Request: Error handling multiple MCP Servers

  #2083 opened Jun 26, 2025

• Test fail with new mistralai version

  #2082 opened Jun 26, 2025

• Feature Request: Support Both JSON & YAML as Tool Output Formats

  #2074 opened Jun 25, 2025

• [Feature Request] Stateful Python Execution Environment for Agents in run_python MCP server (Jupyter-like Workspace)

  #2070 opened Jun 25, 2025

• Feature Request: mcp-run-python support streamable http

  #2059 opened Jun 24, 2025

35 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Improve docs index and README

  #2058 commented on Jun 24, 2025 • 23 new comments

• Add Hugging Face as a provider

  #1911 commented on Jun 26, 2025 • 8 new comments

• deprecate call_id in favor of tool_call_id

  #2028 commented on Jun 27, 2025 • 5 new comments

• Retry prompt model response template

  #2008 commented on Jun 30, 2025 • 5 new comments

• Need custom endpoint support for Bedrock

  #2057 commented on Jun 24, 2025 • 0 new comments

• Add a llama-cpp Model?

  #1801 commented on Jun 28, 2025 • 0 new comments

• Dynamic System prompts not being included on fresh runs with message_history

  #1646 commented on Jun 29, 2025 • 0 new comments

• For list-like `result_models`, LLM tries to call `final_result` multiple times

  #1429 commented on Jun 30, 2025 • 0 new comments

• Gemini causes 'Event loop is closed' when running inside an async context

  #748 commented on Jun 30, 2025 • 0 new comments

• Add media processing settings to GoogleModelSettings

  #2017 commented on Jun 30, 2025 • 0 new comments

• MALFORMED_FUNCTION finishReason in Gemini candidate

  #631 commented on Jun 30, 2025 • 0 new comments

• [OpenAI] Error Handling for Exceeding Maximum Token Limits

  #1098 commented on Jun 30, 2025 • 0 new comments

• Messages passed to history_processor functions include current run messages

  #2050 commented on Jun 30, 2025 • 0 new comments

• add the ability to set a list of allowed tools on an mcp server

  #1219 commented on Jun 30, 2025 • 0 new comments

• Support image output

  #1130 commented on Jun 30, 2025 • 0 new comments

• Add `builtin_tools` to `Agent`

  #1722 commented on Jun 26, 2025 • 0 new comments

• Toolsets

  #2024 commented on Jun 30, 2025 • 0 new comments

• feat: AG-UI adapter

  #2044 commented on Jun 27, 2025 • 0 new comments

• Structured Output Streaming Returning as Single Chunk(Bedrock Anthropic)

  #1889 commented on Jun 24, 2025 • 0 new comments

• Add ability to load conversation history within specified date ranges in CLI

  #1981 commented on Jun 24, 2025 • 0 new comments

• Pre-validation tool response modifications

  #1985 commented on Jun 25, 2025 • 0 new comments

• Add option to filter reasoning content from message history

  #1157 commented on Jun 25, 2025 • 0 new comments

• Show tool response in the running tool span

  #2004 commented on Jun 25, 2025 • 0 new comments

• Create Agents with Oracle Cloud Infrastructure (OCI) Hosted LLMs

  #1385 commented on Jun 25, 2025 • 0 new comments

• Cannot leave a model running

  #1489 commented on Jun 25, 2025 • 0 new comments

• Bedrock: Running agent without tools fails when using messages from an agent with tools

  #1966 commented on Jun 25, 2025 • 0 new comments

• Feature Request: Repeated Evaluations in Evaluation Report

  #2053 commented on Jun 26, 2025 • 0 new comments

• Stop using graphs in agents

  #1976 commented on Jun 26, 2025 • 0 new comments

• Remove deprecated preview models from LatestGeminiModelNames

  #2016 commented on Jun 26, 2025 • 0 new comments

• Pydantic AI V1

  #1974 commented on Jun 26, 2025 • 0 new comments

• Extracting `new_messages()` in the middle of an `Agent.iter` run to support user-in-the-loop approval of tool calls

  #1995 commented on Jun 26, 2025 • 0 new comments

• Access ModelRequest-ModelResponse timestamps and duration as fields

  #1996 commented on Jun 27, 2025 • 0 new comments

• run_stream not working properly with tools in streaming mode

  #1007 commented on Jun 27, 2025 • 0 new comments

• Azure OpenAI API Streaming Response Causes AttributeError in pydantic-ai

  #797 commented on Jun 27, 2025 • 0 new comments

• The reasoning content from deepseek-reasoner is not included in streamed runs but included in non-streamed runs

  #2048 commented on Jun 28, 2025 • 0 new comments