a2aproject/A2A
Agent2Agent (A2A) is an open protocol enabling secure communication and interoperability between diverse generative AI agents to foster collaboration and innovation in the AI ecosystem.
Awesome AI Agents › Multi-Agent Platforms
operand/agency
Agency is a Python library designed to facilitate the creation of agentic systems using the Actor model framework. It provides a minimal yet powerful foundation for developers to build custom agent-based applications that can integrate seamlessly with traditional software systems. The library emphasizes flexibility and scalability, allowing for the development of diverse architectures tailored to specific needs. The core concept in Agency is the 'Agent' class, which represents all entities within the system, including AI-driven agents, software interfaces, and human users. Agents expose 'actions' that other agents can discover and invoke at runtime, enabling dynamic interaction and communication. The API is designed to be straightforward, with class and method-based definitions for agents and their actions. Agency supports concurrency through multiprocessing and multithreading, and it also offers AMQP support for networked agent systems, enhancing performance and scalability. Observability and control are key features, with action and lifecycle callbacks, access policies, permission callbacks, and detailed logging to monitor and manage agent behavior effectively. The library includes a demo application showcasing multiple agent examples, such as OpenAI agents, HuggingFace transformers agents, and operating system access agents. The demo features a Gradio UI and Docker configuration for easy experimentation and development. Agents communicate within a 'Space,' which can be local or networked via AMQP. This design allows for flexible deployment scenarios, from single-application setups to distributed systems. The library is well-documented, with up-to-date resources and examples to assist developers in getting started and exploring advanced features. Overall, Agency is a robust and minimalistic framework aimed at developers interested in experimenting with and building sophisticated agent-based systems, providing the tools necessary for creating scalable, observable, and controllable agent interactions.
https://github.com/operand/agency
Also in Multi-Agent Platforms
IBM/mcp-context-forge
IBM MCP Context Forge is a production-grade Model Context Protocol Gateway that unifies REST and MCP services into a single secure, scalable endpoint for AI clients with federation, virtualization, multi-protocol support, and an optional Admin UI.
a2aproject/a2a-python
The A2A Python SDK is an official toolkit for building and running agentic applications as A2AServers following the Agent2Agent (A2A) Protocol, supporting multiple databases and easy installation.
TooTallNate/proxy-agents
TooTallNate/proxy-agents is a Node.js monorepo containing various HTTP Proxy Agent implementations that enable routing HTTP requests through different proxy protocols.
i-am-bee/acp
ACP is an open protocol enabling seamless communication and coordination between AI agents, applications, and humans through structured multimodal messages.
BigSweetPotatoStudio/HyperChat
HyperChat is an open-source, cross-platform chat client that integrates multiple LLM APIs and the MCP protocol to deliver a versatile and productivity-enhancing chat experience.
ai-boost/awesome-a2a
A curated collection of resources, tools, and implementations for the Google Agent2Agent (A2A) Protocol enabling secure communication and collaboration between AI agents from different vendors and frameworks.
livekit/python-sdks
LiveKit Python SDKs provide real-time video, audio, and data communication capabilities for Python applications, enabling easy integration with LiveKit Cloud or self-hosted servers for building interactive and streaming applications.