2 Releases published by 1 person

• v0.0.15

  published May 15, 2025

• v0.0.16

  published May 21, 2025

29 Pull requests merged by 18 people

• Ensure item.model_dump only contains JSON serializable types

  #801 merged Jun 2, 2025

• Fix #777 by handling MCPCall events in RunImpl

  #799 merged Jun 2, 2025

• Add comment to handoff_occured misspelling

  #792 merged May 30, 2025

• docs: fix typo in docstring for is_strict_json_schema method

  #775 merged May 30, 2025

• Update input_guardrails.py

  #774 merged May 30, 2025

• Fix typo: Replace 'two' with 'three' in /docs/mcp.md

  #757 merged May 30, 2025

• Fix typo in assertion message for handoff function

  #780 merged May 30, 2025

• Small fix for litellm model

  #789 merged May 30, 2025

• Fixed Python syntax

  #665 merged May 29, 2025

• Added RunErrorDetails object for MaxTurnsExceeded exception

  #743 merged May 29, 2025

• Add Portkey AI as a tracing provider

  #785 merged May 29, 2025

• Fix Gemini API content filter handling

  #746 merged May 23, 2025

• Update MCP and tool docs

  #736 merged May 23, 2025

• Fix visualization recursion with cycle detection

  #737 merged May 23, 2025

• fix Gemini token validation issue with LiteLLM

  #735 merged May 21, 2025

• v0.0.16

  #733 merged May 21, 2025

• Add support for local shell, image generator, code interpreter tools

  #732 merged May 21, 2025

• Hosted MCP support

  #731 merged May 21, 2025

• Upgrade openAI sdk version

  #730 merged May 21, 2025

• Add ID generation methods to TraceProvider

  #729 merged May 21, 2025

• DRAFT: Dev/add usage details to Usage class

  #726 merged May 20, 2025

• Add Galileo to external tracing processors list

  #662 merged May 19, 2025

• Added mcp 'instructions' attribute to the server

  #706 merged May 18, 2025

• Create AGENTS.md

  #707 merged May 18, 2025

• v0.0.15

  #701 merged May 15, 2025

• feat: Streamable HTTP support

  #643 merged May 14, 2025

• Update search_agent.py

  #677 merged May 14, 2025

• feat: pass extra_body through to LiteLLM acompletion

  #638 merged May 14, 2025

• Fixed a bug for "detail" attribute in input image

  #685 merged May 14, 2025

16 Pull requests opened by 15 people

• Examples: Fixed agent_patterns/streaming guardrails

  #648 opened May 5, 2025

• feat: Add support for image function tools

  #654 opened May 6, 2025

• Added support for gpt4o-realtime models for Speect to Speech interactions

  #659 opened May 7, 2025

• Create The moon dev

  #709 opened May 17, 2025

• visualize the complete Agent Loop with an interactive UML

  #713 opened May 18, 2025

• Make Runner an abstract base class

  #720 opened May 20, 2025

• docs: fix output guardrail step description

  #724 opened May 20, 2025

• add get_tool_call_name method to extract tool name from ToolCallItem

  #728 opened May 21, 2025

• Jules wip 10459775938551966727

  #739 opened May 22, 2025

• remove "can can" with single can

  #751 opened May 24, 2025

• Add Sessions for Automatic Conversation History Management

  #752 opened May 24, 2025

• Docs: Fix Guardrail definition to include both input and output validations

  #756 opened May 25, 2025

• Fix and Document `parallel_tool_calls` Attribute in ModelSettings

  #763 opened May 26, 2025

• Added support for passing tool_call_id via the RunContextWrapper

  #766 opened May 27, 2025

• grammar fix

  #790 opened May 30, 2025

• Add `exclude_params` feature to `function_tool` for hidden parameter injection

  #795 opened May 30, 2025

50 Issues closed by 13 people

• web_search_preview tool call is not limited by max_turns option

  #783 closed Jun 2, 2025

• Unexpected output type, ignoring: <class 'openai.types.responses.response_output_item.McpCall'>

  #777 closed Jun 2, 2025

• Streaming Output issue got unnecesory text in output

  #753 closed Jun 2, 2025

• How to use openai-agents to call text2image API?

  #725 closed Jun 1, 2025

• How to integrate audio input in agent sdk? (not realtime or voice pipeline, e.g., gpt-4o audio)

  #738 closed Jun 1, 2025

• @function_tool bug?

  #794 closed May 30, 2025

• agents.WebSearchTool bug

  #788 closed May 30, 2025

• Add Portkey AI as tracing Provider in docs

  #786 closed May 29, 2025

• Tracing client error 400 (MCP)

  #710 closed May 29, 2025

• Ba

  #755 closed May 27, 2025

• MCP Stdio- Hosted tools are not supported with the ChatCompletions API

  #761 closed May 26, 2025

• How to pass hardcoded dynamic messages as agent's responses in the chat history ?

  #695 closed May 26, 2025

• litellm proxy with openai-agents

  #696 closed May 26, 2025

• How to pass reasoning=generate_summary for ComputerTool Agent?

  #698 closed May 26, 2025

• Multiple handoffs requested Error tracing platform

  #694 closed May 25, 2025

• Troubleshooting Agent Handoff in Multi-Agent Workflow

  #681 closed May 24, 2025

• Infinite recursion in src/agents/extensions/visualization.py due to circular references

  #668 closed May 23, 2025

• Streaming chain of thoughts

  #721 closed May 22, 2025

• When using Japanese in AzureOpenAI, answers may not be displayed

  #649 closed May 22, 2025

• Threads api

  #674 closed May 22, 2025

• LiteLLM with Gemini token issue

  #734 closed May 21, 2025

• Passing images as input to an agent

  #727 closed May 21, 2025

• llm.txt?

  #670 closed May 20, 2025

• How to make a LiteLLM models run in Reasoning mode

  #671 closed May 20, 2025

• Add an example for telephony voice agent

  #672 closed May 20, 2025

• MCP Server instructions are currently ignored

  #704 closed May 19, 2025

• Does the formatted output (Agent.output_type) require model support?

  #664 closed May 19, 2025

• Question about streaming for subagents and tools, and tool hallucinations

  #667 closed May 19, 2025

• Is there a way to access reasoning_content when calling Runner.run?

  #645 closed May 18, 2025

• Providing a pydantic model instead of docstring for tool parameters.

  #646 closed May 18, 2025

• Creating Agents Dynamically

  #641 closed May 16, 2025

• How to add messages to the conversation history

  #642 closed May 16, 2025

• mlflow.openai.autolog() raises 'NotGiven' object is not iterable when tracing with Ollama + Agents SDK

  #680 closed May 14, 2025

• agents.exceptions.UserError: Hosted tools are not supported with the ChatCompletions API. FGot tool type: <class 'agents.tool.WebSearchTool'>, tool: WebSearchTool(user_location={'type': 'approximate', 'city': '西安'}, search_context_size='medium')

  #194 closed May 14, 2025

• Braicool

  #675 closed May 12, 2025

• Does StopAtTools returns tool result directly to user instead of to LLM?

  #632 closed May 12, 2025

• additionalProperties should not be set for object types

  #608 closed May 11, 2025

• Handoff Agent and Tool Call Not Triggering Reliably in Multi-Agent Setup

  #617 closed May 11, 2025

• How to use on_handoff content in the agent

  #627 closed May 11, 2025

• What is the role of ReasoningItem

  #480 closed May 10, 2025

• example streaming events to the client

  #653 closed May 10, 2025

• Triage agent can not delegate task to handoff agent

  #575 closed May 9, 2025

• Agent gets stuck 'in-progress'

  #647 closed May 8, 2025

• How to use custom LLM Gateway having JWT authetication

  #652 closed May 7, 2025

• Integration of deterministic conversations and other agents

  #603 closed May 6, 2025

• Are MCPServer and MCPServerSse clients?

  #640 closed May 5, 2025

• how to use Code Interpreter or Image Output in OpenAI Agents SDK

  #360 closed May 5, 2025

• ## Custom Model Provider Not Working

  #485 closed May 5, 2025

• function call can not get call_id

  #559 closed May 4, 2025

• How to use llm outputs in the on_handoff function

  #567 closed May 4, 2025

58 Issues opened by 55 people

• User Approval Before Executing MCP Tool in Agent Using mcp_servers

  #802 opened Jun 2, 2025

• Chain-of-thought and structured output

  #800 opened Jun 2, 2025

• Feature Request: Enhanced Run Lifecycle Management - Interrupt and Update Active Runs

  #798 opened Jun 2, 2025

• Is it possible to use both local tools and mcp server tools?

  #797 opened May 31, 2025

• Background thread is created on library import

  #796 opened May 31, 2025

• SDK support for retrieving historical traces

  #793 opened May 30, 2025

• Parallel Tool Calling: Q/Bug?

  #791 opened May 30, 2025

• Processing image/ multi modal responses in function tool results?

  #787 opened May 29, 2025

• [Feature Request] Add REPL-style `run_demo_loop` utility for rapid agent testing and debugging

  #784 opened May 29, 2025

• Rate Limit Support

  #782 opened May 28, 2025

• Misleading Assertion Logic and Message for on_handoff/input_type in `handoff()` function at `agents/handoffs.py`

  #781 opened May 28, 2025

• Return an image from an MCP tool call and pass it to LLM

  #779 opened May 28, 2025

• web_search_preview never triggers agents.lifecycle.AgentHooks.on_tool_start

  #778 opened May 28, 2025

• Always access the OpenAI platform when using the DeepSeek model with LitellmModel

  #773 opened May 28, 2025

• Bug: Pydantic Warnings and `.final_output` Serialization Issue with Code Interpreter File Citations

  #772 opened May 28, 2025

• Only 1 handoff getting called no matter what

  #771 opened May 28, 2025

• function_tool calling - agents orchertratins

  #769 opened May 27, 2025

• Agents Reproducibility - seed ? top_p=0?

  #768 opened May 27, 2025

• How to dynamically add/remove tools in a `tool_use_behavior="run_llm_again"`-loop

  #767 opened May 27, 2025

• Tool calling with LiteLLM and thinking models fail

  #765 opened May 27, 2025

• How to implement handoff between parents and children agents in different paths

  #764 opened May 27, 2025

• parallel_tool_calls default is True in the client, despite docs saying False

  #762 opened May 26, 2025

• ValidationError` from `InputTokensDetails` when using `LitellmModel` with `None` cached\_tokens

  #760 opened May 26, 2025

• LiteLLM + Gemini 2.5 Pro: cached_tokens=None crashes Agents SDK with Pydantic int-validation error

  #758 opened May 26, 2025

• support for bedrock prompt caching

  #750 opened May 24, 2025

• Will session memory be implemented in the future?

  #748 opened May 23, 2025

• Add Session Memory

  #745 opened May 22, 2025

• Gemini API (OpenAI-compatible) `gemini-2.5-flash-preview-05-20` causes `AttributeError` in Agents SDK when `tools` are specified

  #744 opened May 22, 2025

• Issue with Multi-turn Handoff and Slot Filling

  #741 opened May 22, 2025

• Can't implement guardrails with litellm

  #740 opened May 22, 2025

• Unable to use tool calling when using Llama 4 scout via LiteLLM Proxy

  #723 opened May 20, 2025

• How would I handoff a non-reasoning model with tool calls to a reasoning model?

  #722 opened May 20, 2025

• Make intermediate results available when `MaxTurnExceededException` is thrown

  #719 opened May 20, 2025

• Very high response times at random during hand-offs

  #717 opened May 19, 2025

• openai.BadRequestError: Error code: 400 - {'error': {'message': 'Invalid request: prompt is empty or malformed.', 'type': 'invalid_request_error', 'param': None, 'code': None}}

  #716 opened May 19, 2025

• Allow agent to return logprobs

  #715 opened May 19, 2025

• How to pass the control back to triage agent

  #714 opened May 19, 2025

• PyCharm Fails to Recognize Agent Constructor Arguments (v0.0.15)

  #712 opened May 18, 2025

• LLM Cannot Access tool_context.context (e.g. user_uuid) for Tool Input Interpolation in MCP Workflows

  #711 opened May 18, 2025

• Make FuncTool and @function_tool decorated function callable

  #708 opened May 16, 2025

• Custom-LLM <think> tag handling

  #703 opened May 15, 2025

• MCP server restart cause Agent to fail

  #693 opened May 14, 2025

• Feature Request: Support streaming tool call outputs

  #692 opened May 14, 2025

• ImportError: cannot import name 'MCPServerStdio' from 'agents.mcp'

  #691 opened May 14, 2025

• Please add time travel

  #688 opened May 13, 2025

• Agent attempts to use non-existing tool

  #686 opened May 12, 2025

• Feature Request: Allow Separate Models for Tool Execution and Final Response in OpenAI Agent SDK

  #684 opened May 12, 2025

• Add response cost in the Usage

  #683 opened May 12, 2025

• Unable to use reasoning models with tool calls using LitellmModel

  #678 opened May 11, 2025

• How to provide resources with a MCP server?

  #676 opened May 11, 2025

• Error code: 400 "No tool output found for function call"

  #673 opened May 10, 2025

• from agents.extensions.models.litellm_model import LitellmModel

  #666 opened May 8, 2025

• Custom model provider ignored when using agents as tools

  #663 opened May 7, 2025

• First-class streaming tool output

  #661 opened May 7, 2025

• the same isinstance(output, ResponseFunctionToolCall) check twice in "_run_impl.py "

  #658 opened May 7, 2025

• OAuth support for MCPServerSse

  #657 opened May 7, 2025

• Function calling fails on “application/json” MIME type with the latest Gemini models

  #656 opened May 6, 2025

• Input format in agent as tool

  #655 opened May 6, 2025

32 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.

• human-in-the-loop

  #378 commented on May 5, 2025 • 0 new comments

• Tool Calling Running in Loop Until Max-Turn

  #191 commented on May 5, 2025 • 0 new comments

• Streamed Voice Agent Demo - Multiple Performance Issues

  #301 commented on May 6, 2025 • 0 new comments

• Add Intro message function for VoicePipeline

  #488 commented on May 6, 2025 • 0 new comments

• Timeout after 300 seconds with any error message. Could it be rate limiting?

  #511 commented on May 7, 2025 • 0 new comments

• Reasoning model items provide to General model

  #569 commented on May 8, 2025 • 0 new comments

• Retry mechanism for ModelBehaviorError

  #325 commented on May 8, 2025 • 0 new comments

• Missing Handling for `delta.reasoning_content` in `agents.models.chatcmpl_stream_handler.ChatCmplStreamHandler.handle_stream`

  #578 commented on May 9, 2025 • 0 new comments

• Add HTTP Streamable support for MCP's

  #600 commented on May 9, 2025 • 0 new comments

• Websocket streaming audio in realtime from client

  #536 commented on May 12, 2025 • 0 new comments

• History Cleaning

  #545 commented on May 12, 2025 • 0 new comments

• is there a way to block the handoff to an agent based on a custom logic ?

  #585 commented on May 15, 2025 • 0 new comments

• Ordering of events in Runner.run_streamed is incorrect

  #583 commented on May 16, 2025 • 0 new comments

• Question about Tool Call Streaming

  #326 commented on May 16, 2025 • 0 new comments

• Duplicate tool names across MCP servers cause errors

  #464 commented on May 17, 2025 • 0 new comments

• Proper way of managing large context window for ComputerTool

  #111 commented on May 20, 2025 • 0 new comments

• Support for OpenAI agents sdk with Javascript/Typescript

  #240 commented on May 21, 2025 • 0 new comments

• How to make the conversation finally back to the MAIN AGENT

  #527 commented on May 22, 2025 • 0 new comments

• Support for MCP prompts and resources

  #544 commented on May 22, 2025 • 0 new comments

• Add Support for Image Return in Agent Tools

  #341 commented on May 24, 2025 • 0 new comments

• Cannot get the last tool_call_output event in stream_events when MaxTurnsExceeded

  #526 commented on May 26, 2025 • 0 new comments

• Enhance `on_tool_start` Hook to Include Tool Call Arguments

  #252 commented on May 26, 2025 • 0 new comments

• Best practices for long running tools?

  #295 commented on May 29, 2025 • 0 new comments

• Why does the Computer protocol not have the goto method?

  #547 commented on May 29, 2025 • 0 new comments

• [Bug]: SDK crashes when `choices` is `None` (provider-error payload)

  #604 commented on May 30, 2025 • 0 new comments

• Human-In-The-Loop Architecture should be implemented on top priority!

  #636 commented on Jun 2, 2025 • 0 new comments

• Add tool call parameters for `on_tool_start` hook

  #253 commented on May 23, 2025 • 0 new comments

• add reasoning content to ChatCompletions

  #494 commented on May 5, 2025 • 0 new comments

• Make input/new items available in the run context

  #572 commented on May 13, 2025 • 0 new comments

• Add a new GH Actions job to automatically update translated document pagse

  #598 commented on May 21, 2025 • 0 new comments

• [MCP][Utils] Add support for FastMCP processing

  #631 commented on May 20, 2025 • 0 new comments

• fix: add ensure_ascii=False to json.dumps for correct Unicode output

  #639 commented on May 25, 2025 • 0 new comments