6 Releases published by 2 people

• v0.0.9

  published Apr 7, 2025

• v0.0.10

  published Apr 15, 2025

• v0.0.11

  published Apr 15, 2025

• v0.0.12

  published Apr 22, 2025

• v0.0.13

  published Apr 24, 2025

• v0.0.14 v0.0.14

  published Apr 30, 2025

55 Pull requests merged by 17 people

• 0.0.14 release

  #635 merged Apr 30, 2025

• Update litellm version

  #626 merged Apr 29, 2025

• docs: add FutureAGI to tracing documentation

  #592 merged Apr 25, 2025

• Make the TTS voices type exportable

  #577 merged Apr 24, 2025

• Add usage to context in streaming

  #595 merged Apr 24, 2025

• v0.0.13

  #593 merged Apr 24, 2025

• More tests for cancelling streamed run

  #590 merged Apr 24, 2025

• Fix stream error using LiteLLM

  #589 merged Apr 24, 2025

• Prevent MCP ClientSession hang

  #580 merged Apr 24, 2025

• Create to_json_dict for ModelSettings

  #582 merged Apr 24, 2025

• Allow cancel out of the streaming result

  #579 merged Apr 23, 2025

• Examples: Fix financial_research_agent instructions

  #573 merged Apr 23, 2025

• Adding extra_headers parameters to ModelSettings

  #550 merged Apr 23, 2025

• v0.0.12

  #564 merged Apr 22, 2025

• Pass through organization/project headers to tracing backend, fix speech_group enum

  #562 merged Apr 21, 2025

• Docs and tests for litellm

  #561 merged Apr 21, 2025

• RFC: automatically use litellm if possible

  #534 merged Apr 21, 2025

• Fix visualize graph filename to without extension.

  #554 merged Apr 21, 2025

• Start and finish streaming trace in impl metod

  #540 merged Apr 21, 2025

• Enable non-strict output types

  #539 merged Apr 21, 2025

• Examples for image inputs

  #553 merged Apr 21, 2025

• Docs: Switch to o3 model; exclude translated pages from search

  #533 merged Apr 17, 2025

• Docs for LiteLLM integration

  #532 merged Apr 16, 2025

• Litellm integration

  #524 merged Apr 16, 2025

• Show repo name/data in docs

  #525 merged Apr 16, 2025

• Extract chat completions streaming helpers

  #523 merged Apr 15, 2025

• Extract chat completions conversion code into helper

  #522 merged Apr 15, 2025

• Run CI on all commits, not just ones on main

  #521 merged Apr 15, 2025

• v0.0.11

  #520 merged Apr 15, 2025

• Only include stream_options when streaming

  #519 merged Apr 15, 2025

• Examples and tests for previous_response_id

  #512 merged Apr 15, 2025

• Support parallel_tool_calls=False for ChatCompletion API

  #513 merged Apr 15, 2025

• v0.0.10

  #514 merged Apr 15, 2025

• Evolve the doc translation workflow by using gpt-4.1

  #507 merged Apr 15, 2025

• Previous response id

  #509 merged Apr 15, 2025

• Move dev-only deps to dev section

  #506 merged Apr 15, 2025

• Replace referencable_id with response_id

  #508 merged Apr 15, 2025

• Fix type annotations examples

  #496 merged Apr 14, 2025

• Don't run tracing shutdown behavior if disabled

  #503 merged Apr 14, 2025

• Example for streaming guardrails

  #505 merged Apr 14, 2025

• Add docs for customizng agent-as-tool

  #504 merged Apr 14, 2025

• Fix typos and misspellings

  #486 merged Apr 14, 2025

• Add extra request parameters extra_query and extra_body

  #500 merged Apr 14, 2025

• fix: ensure metadata is non-null in response handling

  #483 merged Apr 11, 2025

• Add model_settings guide to models document page

  #484 merged Apr 11, 2025

• Improve translation pipeline details

  #475 merged Apr 10, 2025

• add overwrite mechanism for stream_options

  #465 merged Apr 10, 2025

• Adding scope and span_data doc strings

  #463 merged Apr 9, 2025

• examples(research_bot): Add space to INSTRUCTIONS

  #439 merged Apr 8, 2025

• docs: Make default model clear

  #457 merged Apr 8, 2025

• Update computer-use example to support simultaneous key presses

  #452 merged Apr 8, 2025

• Add i18n support to the documents

  #460 merged Apr 8, 2025

• v0.0.9

  #456 merged Apr 7, 2025

• Don't send the "store" param unless its hitting OpenAI

  #455 merged Apr 7, 2025

• Ensure MCP works when inputSchema.properties is missing

  #454 merged Apr 7, 2025

20 Pull requests opened by 18 people

• add reasoning content to ChatCompletions

  #494 opened Apr 14, 2025

• Added cached_tokens to the usage monitoring.

  #555 opened Apr 20, 2025

• Feature/unified agent endpoint

  #558 opened Apr 21, 2025

• Add File Loading Utilities for Agent Instructions

  #565 opened Apr 22, 2025

• Make input/new items available in the run context

  #572 opened Apr 22, 2025

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

  #598 opened Apr 24, 2025

• Update README.md

  #606 opened Apr 26, 2025

• Remove vscode settings from the project

  #607 opened Apr 26, 2025

• Add ProviderError exception and handle missing LLM response

  #609 opened Apr 26, 2025

• Fix spacing inconsistencies and adhere to style guidelines

  #612 opened Apr 27, 2025

• feat: use stream api only

  #629 opened Apr 29, 2025

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

  #631 opened Apr 30, 2025

• feat: patch prompt for models only support json-mode

  #633 opened Apr 30, 2025

• feat: Implement get_tool_call_output method in RunResultBase and update doc

  #637 opened May 1, 2025

• feat: pass extra_body through to LiteLLM acompletion

  #638 opened May 1, 2025

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

  #639 opened May 2, 2025

• feat: Streamable HTTP support

  #643 opened May 2, 2025

• [fix] use openai model provider as default

  #644 opened May 3, 2025

• Examples: Fixed agent_patterns/streaming guardrails

  #648 opened May 5, 2025

• Fix typos in documentation and event naming across multiple files

  #651 opened May 6, 2025

110 Issues closed by 31 people

• 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

• Tools should not be exeucted until all input guardrails have completed

  #624 closed May 2, 2025

• Files in the input user prompt

  #557 closed May 2, 2025

• from agents.extensions.models.litellm_model import LitellmModel

  #621 closed May 1, 2025

• Accessing reasoning tokens of another llm model in agents sdk

  #462 closed May 1, 2025

• 'AsyncOpenAI' Object Has No Attribute 'responses' in Async Runner & 'this event loop is already running' in Sync Runner

  #542 closed May 1, 2025

• [Bug]: ModuleNotFoundError: No module named 'enterprise' When Using litellm==1.48.1 in Google Colab

  #614 closed Apr 30, 2025

• ModuleNotFoundError: No module named 'enterprise' #10353

  #613 closed Apr 30, 2025

• Add HTTP (non-stdio) MCP server support to Agents SDK

  #616 closed Apr 29, 2025

• OpenAI Agents SDK unable to contact local endpoint hosted by Ollama / LM Studio

  #625 closed Apr 29, 2025

• https://static.hotmart.com/checkout/widget.min.js

  #619 closed Apr 29, 2025

• AWS Bedrock via LiteLLM

  #620 closed Apr 29, 2025

• how to print Mcp tools print

  #615 closed Apr 28, 2025

• Is RunContext thread safe?

  #537 closed Apr 28, 2025

• agent output format and handoff stability

  #568 closed Apr 27, 2025

• Optimize Latency for Parallel Agent Runs with Streaming

  #498 closed Apr 27, 2025

• How would you handle Pydantic output_type validation and retries?

  #530 closed Apr 27, 2025

• handoff() returned Handoff objects aren’t recognized by Agent(handoffs=…) without manual .name alias

  #599 closed Apr 27, 2025

• Zero Data Retention Organizations are incompatible with the Responses API/OAI State Management/Tool Calls

  #605 closed Apr 26, 2025

• Error in Stream in Runner.run_streamed() with LitellmModel(Model) class

  #601 closed Apr 26, 2025

• Support AWS Bedrock

  #86 closed Apr 26, 2025

• Issue when processing real time audio from a Twilio media stream

  #304 closed Apr 26, 2025

• what is the best practices to have faster voices conversation?

  #306 closed Apr 26, 2025

• AttributeError: module 'planner_agent' has no attribute 'handoffs'

  #596 closed Apr 25, 2025

• It seems I'm unable to access the file `sample.txt`

  #584 closed Apr 25, 2025

• Optional Fields in Agent Output Cause JSON Schema Error with AWS Bedrock

  #586 closed Apr 25, 2025

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

  #585 closed Apr 25, 2025

• 'AgentOutputSchema' object has no attribute '__mro__'

  #597 closed Apr 25, 2025

• How can I pass dynamic instruction to the agent

  #482 closed Apr 25, 2025

• context.usage returns 0 in streaming mode

  #594 closed Apr 24, 2025

• LiteLLM extension crashes with run_streamed

  #587 closed Apr 24, 2025

• Returning function call's responses in `raw_response_event`

  #328 closed Apr 24, 2025

• Canceling the stream from result.stream_events()

  #574 closed Apr 23, 2025

• Why are tech contributions ignored and reimplemented without discussion and/or consensus, shall we simply stop 🛑 contributing back?

  #566 closed Apr 23, 2025

• When will the .NET version be available?

  #571 closed Apr 23, 2025

• How to make hand-off decisions more reliable?

  #541 closed Apr 23, 2025

• Nested objects in function tool input are empty

  #563 closed Apr 22, 2025

• openai-agents package ot installing correctly. "ModuleNotFound agents"

  #399 closed Apr 22, 2025

• Missing packages on install for voice with macOS Intel

  #478 closed Apr 22, 2025

• Adding Langsmith trace processor introduces huge latency to chat

  #529 closed Apr 22, 2025

• ValueError when finishing trace in Runner.run_streamed(): “token was created in a different Context”

  #538 closed Apr 21, 2025

• StreamingResponse with FastAPI return error ContextVar

  #435 closed Apr 21, 2025

• how to use custom json schema for agent's output?

  #528 closed Apr 21, 2025

• Incorrect package name in documentation: 'pip install openai-agents' should be 'openai_agents'

  #560 closed Apr 21, 2025

• function_call can not work

  #467 closed Apr 21, 2025

• MCP server with custom OpenAI client

  #470 closed Apr 21, 2025

• multi-agent use handoffs

  #471 closed Apr 21, 2025

• 多智能体协作完成一个复杂的任务

  #473 closed Apr 21, 2025

• When the agent performs a handoff, can I append the new user input to the end of the handoff history log?

  #408 closed Apr 20, 2025

• Fact Checking Guardrails

  #416 closed Apr 20, 2025

• Expose max_turns Parameter When Running an Agent as a Tool

  #551 closed Apr 20, 2025

• [non-fatal] Tracing client error 401, Incorrect API key provided

  #535 closed Apr 17, 2025

• MCP with context

  #411 closed Apr 17, 2025

• why?

  #441 closed Apr 17, 2025

• UTF-8 encoding issues in tool calling params with GPT-4o model

  #315 closed Apr 16, 2025

• context management between the agents in a multi agent set up

  #348 closed Apr 16, 2025

• timeout not work for post_writer in mcp

  #517 closed Apr 16, 2025

• Model stream error when running in Docker

  #518 closed Apr 15, 2025

• Why full history as input is still necessary?

  #102 closed Apr 15, 2025

• Runner hangs when using MCPServerStdio after successful tools/list

  #434 closed Apr 15, 2025

• RunResultStreaming.stream_events() raises ValueError when reset_current=True in async environments

  #515 closed Apr 15, 2025

• How to set max_turn for agents set as tools.

  #497 closed Apr 15, 2025

• Running Agents in Parallel and Traces

  #481 closed Apr 15, 2025

• Set `max_turns` for agent as tool call

  #448 closed Apr 15, 2025

• Is accessing _trace the proper way of getting default trace ID?

  #447 closed Apr 15, 2025

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

  #420 closed Apr 15, 2025

• How do you use prompt walk-around function calling with this SDK to improve Agent performance on o3-mini?

  #491 closed Apr 15, 2025

• Clarify Python Version Compatibility

  #490 closed Apr 14, 2025

• Output Guardrails while Streaming

  #495 closed Apr 14, 2025

• OPENAI_API_KEY is not set, skipping trace export

  #190 closed Apr 14, 2025

• store=true does not work for Responses API

  #175 closed Apr 14, 2025

• Agent as tool does not keep the run config

  #476 closed Apr 14, 2025

• The model `computer-use-preview-2025-03-11` does not exist or you do not have access to it.

  #458 closed Apr 14, 2025

• Add the possibility to add extra body fields in the RunConfig or Agents

  #487 closed Apr 14, 2025

• Multiple handoffs requested error

  #499 closed Apr 14, 2025

• ValueError during agents.tracing shutdown even when tracing is disabled

  #502 closed Apr 14, 2025

• tool_choice error

  #468 closed Apr 14, 2025

• Performance Issue with Agents - Synchronous API Call

  #389 closed Apr 14, 2025

• How do we use `stream_events()` when `tool_choice` = "function_name" ?

  #390 closed Apr 14, 2025

• MCP Prompts API support

  #396 closed Apr 14, 2025

• I really can't figure out how to invoke multiple MCP clients and use handoffs. The official documentation doesn't provide any examples to support this either.

  #425 closed Apr 14, 2025

• Why the Agent SDK doesn't need the "description schema" of a tool

  #427 closed Apr 14, 2025

• Distributed multi agent deployment using handoffs

  #406 closed Apr 13, 2025

• Modify tool invocation results.

  #407 closed Apr 13, 2025

• Make tools as output of an agent

  #409 closed Apr 13, 2025

• Long response times when using multiple agents with agents_as_tools

  #421 closed Apr 13, 2025

• Is there a way to force handoffs?

  #479 closed Apr 11, 2025

• Availability on Azure

  #379 closed Apr 11, 2025

• AttributeError: module 'tensorflow' has no attribute 'contrib'

  #469 closed Apr 10, 2025

• pip install OpenAI-Agent.

  #477 closed Apr 10, 2025

• Agents SDK output_type Generates Invalid JSON Schema for API response_format w Unions

  #474 closed Apr 10, 2025

• Installation fails on Alpine (musl-based) Docker images — missing wheels or source build fails?

  #466 closed Apr 10, 2025

• How to trans sdk output to chat completions

  #393 closed Apr 10, 2025

• How to get RunContextWrapper fields from Run.run_streamed ?

  #453 closed Apr 9, 2025

• How to use Handoff and MCP together？

  #459 closed Apr 9, 2025

• Documentation and implementation mismatch for default model in Agent class

  #440 closed Apr 8, 2025

• TTS starts after entire text is received

  #384 closed Apr 8, 2025

• 400 BadRequestError with Gemini provider – Invalid JSON payload ("metadata" and "store" fields)

  #443 closed Apr 8, 2025

• /agents/mcp/server.py", line 115, in list_tools raise UserError("Server not initialized. Make sure you call `connect()` first.") agents.exceptions.UserError: Server not initialized. Make sure you call `connect()` first.

  #446 closed Apr 7, 2025

• OpenAI Agents SDK calls OpenAI API with invalid schema when MCP tool definitions have no `properties` field

  #449 closed Apr 7, 2025

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

  #413 closed Apr 7, 2025

• The depth agents handsoff draw_graph function can auto find out?

  #359 closed Apr 7, 2025

• windows run mcp error：

  #363 closed Apr 7, 2025

• How to implement tool calls and outputs flow between backend agents and client similar to Assistant API?

  #377 closed Apr 7, 2025

• Support for function tool direct/structured output

  #436 closed Apr 6, 2025

48 Issues opened by 42 people

• How to use custom LLM Gateway having JWT authetication

  #652 opened May 6, 2025

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

  #649 opened May 5, 2025

• Agent gets stuck 'in-progress'

  #647 opened May 5, 2025

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

  #646 opened May 5, 2025

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

  #645 opened May 4, 2025

• How to add messages to the conversation history

  #642 opened May 2, 2025

• Creating Agents Dynamically

  #641 opened May 2, 2025

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

  #636 opened May 1, 2025

• ValueError: <Token var=<ContextVar name='current_trace' default=None at 0x7f9c4462cf40> at 0x7f9c42267c80> was created in a different Context

  #634 opened Apr 30, 2025

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

  #632 opened Apr 30, 2025

• no attribute error occurs while calling MCP

  #630 opened Apr 30, 2025

• Intent Classifier Support

  #628 opened Apr 29, 2025

• How to use on_handoff content in the agent

  #627 opened Apr 29, 2025

• on_agent_start hook should be more performant

  #623 opened Apr 29, 2025

• Can we use agent.run instead of Runner.run(starting_agent=agent)

  #622 opened Apr 29, 2025

• Resource tracker warning (leaked semaphores) with MCPServerStdio

  #618 opened Apr 28, 2025

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

  #617 opened Apr 28, 2025

• Bug: style guideline and formatting inconsistencies

  #611 opened Apr 27, 2025

• [Bug]: UnicodeDecodeError when importing litellm_model on Windows

  #610 opened Apr 26, 2025

• additionalProperties should not be set for object types

  #608 opened Apr 26, 2025

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

  #604 opened Apr 25, 2025

• Add HTTP Streamable support for MCP's

  #600 opened Apr 25, 2025

• openai-agents-dotnet

  #588 opened Apr 24, 2025

• Ordering of events in Runner.run_streamed is incorrect

  #583 opened Apr 24, 2025

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

  #578 opened Apr 23, 2025

• input_guardrail is skipped

  #576 opened Apr 23, 2025

• Triage agent can not delegate task to handoff agent

  #575 opened Apr 23, 2025

• bugs in run.py

  #570 opened Apr 22, 2025

• Reasoning model items provide to General model

  #569 opened Apr 22, 2025

• Add llms.txt in the documentation

  #556 opened Apr 20, 2025

• Add the possibility to add extra header fields in the RunConfig or Agents

  #549 opened Apr 19, 2025

• Usage tokens no longer automatically show

  #548 opened Apr 18, 2025

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

  #547 opened Apr 18, 2025

• History Cleaning

  #545 opened Apr 18, 2025

• Support for MCP prompts and resources

  #544 opened Apr 18, 2025

• Websocket streaming audio in realtime from client

  #536 opened Apr 17, 2025

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

  #527 opened Apr 16, 2025

• Cannot get the last tool_call_output event in stream_events when MaxTurnsExceeded

  #526 opened Apr 16, 2025

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

  #511 opened Apr 15, 2025

• Different Traces View at OpenAI Traces for Responses vs ChatCompletions

  #501 opened Apr 14, 2025

• Obvious Pauses Between Text Segments in Current `OpenAITTSModel` Implementation Affect Speech Fluency

  #493 opened Apr 14, 2025

• Add reasoning support for custom models.

  #492 opened Apr 13, 2025

• Automatic Audio Feedback on Silence in VoicePipeline

  #489 opened Apr 13, 2025

• Add Intro message function for VoicePipeline

  #488 opened Apr 13, 2025

• What is the role of ReasoningItem

  #480 opened Apr 11, 2025

• A2A (Agent2Agent) support

  #472 opened Apr 10, 2025

• Duplicate tool names across MCP servers cause errors

  #464 opened Apr 9, 2025

• Support hosted tools across different model providers

  #461 opened Apr 8, 2025

21 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.

• Add tool call parameters for `on_tool_start` hook

  #253 commented on Apr 29, 2025 • 1 new comment

• chore: add missing openai.fm model names to `TTSModelSettings.voice` type

  #269 commented on Apr 15, 2025 • 0 new comments

• Fix typos in event names and comments across multiple files

  #155 commented on Apr 11, 2025 • 0 new comments

• Streamed Voice Agent Demo - Multiple Performance Issues

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

• Tool Calling Running in Loop Until Max-Turn

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

• human-in-the-loop

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

• Retry mechanism for ModelBehaviorError

  #325 commented on May 2, 2025 • 0 new comments

• invalid_request_error when using "chat_completions" with triage agent (gemini -> any other model)

  #237 commented on May 1, 2025 • 0 new comments

• Support for OpenAI agents sdk with Javascript/Typescript

  #240 commented on Apr 30, 2025 • 0 new comments

• Random transcript gets printed/generated when talking to the voice agent implemented using "VoicePipline" . Eg - "Transcription: Kurs." Mind you there is no background noise.

  #368 commented on Apr 29, 2025 • 0 new comments

• Support For CodeAct In The Future?

  #383 commented on Apr 29, 2025 • 0 new comments

• 'handoffs' and 'agent.as_tool' have different performances.

  #224 commented on Apr 29, 2025 • 0 new comments

• Enhance `on_tool_start` Hook to Include Tool Call Arguments

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

• Add `reasoning_content` to ChatCompletions

  #415 commented on Apr 22, 2025 • 0 new comments

• Golang implementation / alternative

  #350 commented on Apr 15, 2025 • 0 new comments

• SSL error for streaming voice terminal app

  #423 commented on Apr 14, 2025 • 0 new comments

• Tool calling with Gemini + structured_output

  #236 commented on Apr 13, 2025 • 0 new comments

• Local memory and context window length

  #177 commented on Apr 13, 2025 • 0 new comments

• Question about Tool Call Streaming

  #326 commented on Apr 10, 2025 • 0 new comments

• Streaming Fails Due to `{"include_usage": True}`

  #442 commented on Apr 10, 2025 • 0 new comments

• will there be a Java version of openai-agents-java in the future?

  #115 commented on Apr 8, 2025 • 0 new comments