Highlights

• New integration
  • PortKey
  • Mem0
  • Cerebras
• Improved support of Ollama and OpenAI API key.

Thanks to@Hk669@marklysze@yiranwu0@henrytwo@LaraNonino@randombet and all the other contributors!

What's Changed

• Use utf-8 encoding at subprocess.run() by@Hk669 in#23
• moved the package in notebooks from pyautogen to autogen by@Hk669 in#34
• Portkey Integration with Autogen by@marklysze in#38
• Fix Exception by@yiranwu0 in#21
• PR 3362 - fix ImportFromModule by@marklysze in#42
• Ecosystem Mem0 by@marklysze in#40
• Ability to add MessageTransforms to the GroupChat's Select Speaker nested chat (speaker_selection_method='auto') by@marklysze in#2
• Update usage of built-in LLMs in Microsoft Fabric by@marklysze in#41
• OpenAI API key format validation by@marklysze in#44
• Cerebras Integration by@henrytwo in#48
• Cerebras - streaming fix and moving exception handling by@marklysze in#51
• Fix formatting of Cerebras docs by@henrytwo in#52
• Fix: added agent name to client.create calls to prevent runtime loggi… by@LaraNonino in#47
• Updated parameter handling for Ollama by@marklysze in#45
• bump version to 0.3.1 by@sonichi in#53
• Fix.devcontainer by@Hk669 in#56
• [Falkor DB Integration Part 1] Implement FalkorGraphQueryEngine by@randombet in#37

New Contributors

• @henrytwo made their first contribution in#48
• @LaraNonino made their first contribution in#47

Full Changelog:v0.3.0...v0.3.1

Contributors

This is a major release since v0.2.0, containing 35 minor releases (from v0.2.0 to v0.2.35).

Pypi package name: Starting from this release, we will start using theautogen PyPI package. We extend our sincere gratitude to the original owner ofautogen PyPI package for generously transferring it to us.

License: Starting from this release, our project starts using the Apache 2.0 license. Please refer to this notice for details:https://github.com/autogen-ai/autogen/blob/main/NOTICE.md. This enhances our commitment to open-source collaboration while providing additional protections for contributors and users alike.

Highlights since v0.2.0

🚀 Expanded Model Support

• Improved integration with non-OpenAI models: added support for Gemini, Groq, Cohere, Anthropic (including Claude 3), Mistral, and Together.AI models:https://autogen-ai.github.io/autogen/blog/2024/06/24/AltModels-Classes

🤖 New Agent Capabilities

• Introduced OpenAI assistant API support:https://autogen-ai.github.io/autogen/docs/notebooks?tags=OpenAI+Assistant
  Enhanced multimodal capabilities, including vision for ConversableAgents:https://autogen-ai.github.io/autogen/docs/notebooks?tags=multimodal
• Generalized teachability and extensibility: Allow any conversable agent to be teachable via a new and generic extensibility mechanism. Example:https://autogen-ai.github.io/autogen/docs/notebooks?tags=teachability
• Added AgentOptimizer for training LLM agents:https://autogen-ai.github.io/autogen/blog/2023/12/23/AgentOptimizer

💬 Advanced Conversation Programming

• Finite State Machine (FSM) based group chat for fine-grained control:https://autogen-ai.github.io/autogen/docs/notebooks/agentchat_groupchat_finite_state_machine/
• Sequential chats:https://autogen-ai.github.io/autogen/docs/notebooks?tags=sequential+chats
• Nested chats:https://autogen-ai.github.io/autogen/docs/notebooks?tags=nested+chat
• Enhanced async support for multi-chats:https://autogen-ai.github.io/autogen/docs/notebooks?tags=async

🔍 Improved RAG and Retrieval

• Added support for Qdrant, MongoDB, and Azure Cognitive Search:https://autogen-ai.github.io/autogen/docs/notebooks?tags=RAG
• Implemented customizable vector databases for retrieval-augmented chat

🛠️ Enhanced Tool and Function Calling

• Improved function calling support:https://autogen-ai.github.io/autogen/docs/notebooks?tags=tool%2Ffunction
• Introduced "Function Inception" for dynamic function updates during conversations:https://github.com/autogen-ai/autogen/blob/main/notebook/agentchat_inception_function.ipynb

💻 Code Execution and Development

• Enhanced code executors, including Docker-based and stateful executors:https://autogen-ai.github.io/autogen/docs/tutorial/code-executors
• Added support for HTML, CSS, and Javascript in LocalCommandLineCodeExecutor

🆕 New Features and Integrations

• Implemented TransformMessage for improved long context handling:https://autogen-ai.github.io/autogen/docs/notebooks/agentchat_transform_messages/
• Added runtime logging integration with AgentOps:https://autogen-ai.github.io/autogen/blog/2024/07/25/AgentOps
• Introduced AutoGenBench for agent evaluation:https://autogen-ai.github.io/autogen/blog/2024/01/25/AutoGenBench
• Created AutoGen Studio (formerly AutoGen Assistant) with new features:https://autogen-ai.github.io/autogen/blog/2023/12/01/AutoGenStudio

📚 Expanded Documentation and Examples

• Added comprehensive user guides, tutorials, and notebooks:
  • Tutorial:https://autogen-ai.github.io/autogen/docs/tutorial
  • Guides:https://autogen-ai.github.io/autogen/docs/topics
  • Notebooks:https://autogen-ai.github.io/autogen/docs/notebooks
• Expanded the gallery of example applications:https://autogen-ai.github.io/autogen/docs/Gallery

⚡ Performance and Usability Improvements

• Implemented caching improvements, including Redis and diskcache support
• Enhanced group chat functionality and speaker selection
• Improved cost calculation and usage tracking

🌐 New Language Support

• Introduced AutoGen.NET for .NET developers

We extend the warmest gratitude to all the contributors involved, including@sonichi@qingyun-wu@marklysze@skzhang1@LinxinS97@JieyuZ2@Hk669@kevin666aa@BeibinLi@thinkall@ekzhu@jackgerrits@afourney@davorrunje@yenif@rickyloynd-microsoft@gagb@victordibia@olgavrou@LittleLittleCloud@lalo@WaelKarkoub@joshkyh@amihalik
@rajan-chari@areibman@yiranwu0@julianakiseleva@randombet@cheng-tan@elecnix@IANTHEREAL@ShobhitVishnoi30@olaoluwasalami@hodgesmr@cpacker@PyroGenesis@axa99@momuno@Anush008,@ranfysvalle02,@joris-swapfiets@vijaykramesh@balakreshnan@bitnom@chengxuan233@tyler-suard-parker@kittyandrew@btcioner@yuandong-tian@maxim-saplin,@r3d91ll,@Yanni8,@shhdgit,@ashutoshpw@cool-RR@dragosMC91@danyrouh@rlam3@bitsnaps@GregorD1A1@byronxu99@cemtu@olimoz@jtrugman@namanbarkiya@eltociear@pmalarme@freedeaths@pretherford@samuelogboye@Xtrah@abhaymathur21@mrwadams@yousonnet@bobbravo2@FarshidShafia@gunnarku@AaronWard@dkirsche@RohitRathore1@shreyas36@KazooTTT@swiecki@truebit@bassmang@bmuskalla@andreyseas@schauppi@Dave2011@christianladron@huskydoge@panckreous@veh3546@jamesliu@shouldnotappearcalm@Wannabeasmartguy@XHMY@jtoy@Andrew8xx8@benstein@sharsha315@levscaut@Karthikeya-Meesala@r-b-g-b@martinzrrl@sethupavan12@wmwxwa@JoshTrim@giorgossideris@rihp@Mai0313@DustinX@lordlinus@saberuster@Knucklessg1@luxuncang@asandez1@themataleao@beyonddream@ginward@gbrvalerio@DavidLuong98@erezak@zbram101@r4881t@robraux@thongonary@moresearch@shippy@ACHultman [@Gr3atWh173](https://github.com/...

Contributors

What's Changed

• update repo doc by@qingyun-wu in#7
• Updated development instructions inline with changed URLs by@marklysze in#8
• Web re-deploy by@skzhang1 in#10
• fixing urls by@skzhang1 in#11
• Updated repo links by@marklysze in#12
• Integrate Mem0 for providing long-term memory for AI Agents by@Dev-Khant in#9
• update notebooks by@sonichi in#16
• update more URLs by@skzhang1 in#17
• update ecosystem and dotnet docs by@Hk669 in#18
• Update license to Apache License Version 2.0 by@qingyun-wu in#19
• Update Image and website url by@yiranwu0 in#14
• Improve doc: update optional installation dependencies by@yiranwu0 in#13
• Amazon Bedrock Client for AutoGen by@marklysze in#6
• Group Chat - Updates to resume by@marklysze in#20
• [Graph RAG] Init Commit with GraphRag interfaces by@randombet in#1
• Fix Anthropic system messages by@marklysze in#5
• Ollama Client (with tool calling) by@marklysze in#3
• Transform to add an agent's name into the message content by@marklysze in#4
• agent builder tests by@qingyun-wu in#27
• Version and pypi package by@qingyun-wu in#28
• license format by@qingyun-wu in#29
• Package by@qingyun-wu in#30
• update version by@qingyun-wu in#31

New Contributors

• @qingyun-wu made their first contribution in#7
• @marklysze made their first contribution in#8
• @skzhang1 made their first contribution in#10
• @Dev-Khant made their first contribution in#9
• @sonichi made their first contribution in#16
• @Hk669 made their first contribution in#18
• @yiranwu0 made their first contribution in#14
• @randombet made their first contribution in#1

Full Changelog:v0.2.35...v0.3.0b4

Contributors

Highlights

• Enhanced Mistral client support
• Enhanced RAG support
• Enhanced message formatting

Thanks to@WaelKarkoub,@cbelwal,@HenryKobin,@LittleLittleCloud,@olaoluwasalami,@realmorrisliu,@marklysze and all contributors.

What's Changed

• Removes Support ForTransformChatHistory andCompressibleAgent by@WaelKarkoub inmicrosoft#3313
• Updated Program.cs for Autogen.BasicSample to give menu driven options by@cbelwal inmicrosoft#3346
• Remove dependency on RetrieveAssistantAgent for RetrieveChat by@thinkall inmicrosoft#3320
• Missing backticks breaking documentation in groupchat.last_speaker by@HenryKobin inmicrosoft#3357
• Update Mistral client class to support new Mistral v1.0.1 package by@marklysze inmicrosoft#3356
• adding a new page Function comparison between Python AutoGen and Auto… by@olaoluwasalami inmicrosoft#3367
• Update README.md to use camera-ready by@gagb inmicrosoft#3370
• Add OpenAI Gemini Example for VertexAI Notebook by@luxzoli inmicrosoft#3290
• [.NET] Add cache control to Anthropic Client by@DavidLuong98 inmicrosoft#3372
• Correct documentation: Replace NEVER with TERMINATE in human-in-the-loop.ipynb by@realmorrisliu inmicrosoft#3379
• [.Net] Dispose kernel after running dotnet interactive tests by@LittleLittleCloud inmicrosoft#3378
• Ensure 'name' on initial message by@marklysze inmicrosoft#2635
• [.Net] Enable step-by-step execution for two-agent conversation SendAsync API by@LittleLittleCloud inmicrosoft#3360

New Contributors

• @cbelwal made their first contribution inmicrosoft#3346
• @HenryKobin made their first contribution inmicrosoft#3357
• @realmorrisliu made their first contribution inmicrosoft#3379

Full Changelog:microsoft/autogen@v0.2.34...v0.2.35

Contributors