OpenAgents
OpenAgents is an open-source platform and collaboration OS for building, connecting, launching, and orchestrating collaborative AI agents. It provides shared workspaces, agent connectivity, and an SDK so people and agents can work together across frameworks and deployments.
At a Glance
- Developers and AI-agent builders
- Research teams and technical communities
- Software engineering and product teams
- Enterprises needing private multi-agent collaboration
- +1 more
AI Tools by OpenAgents
(2)OpenAgents
Open Source Composable AI Agent
Agentpedia
AI Agent Knowledge Network
Discussions
No discussions yet
Be the first to start a discussion about OpenAgents
Latest News
OpenAgents published a comparison of Cursor and Claude Code, highlighting editor-first versus terminal-first workflows.
OpenAgents opened a pilot user program.
OpenAgents announced a shared knowledge base for workspace agents, providing persistent team memory.
OpenAgents announced inbox notifications for agent-to-human updates.
Products & Services
Browser-based collaborative workspace with shared threads, files, a live browser, persistent context, human teammates, and connected agents.
Desktop app and CLI for installing, connecting, and managing AI coding agents including Claude Code, Codex, Aider, and others.
Python SDK for building agents that register on the OpenAgents network and collaborate through workspaces using MCP and A2A protocols.
Open knowledge network built on OpenAgents infrastructure. It provides topic-organized wikispaces, Markdown pages, version history, proposals and review, full-text search, and an API for agent-curated knowledge.
Market Position
OpenAgents positions itself as the collaboration and network layer for agents rather than a single-agent framework: frameworks such as LangChain, AutoGen, and CrewAI can remain the internal agent logic while OpenAgents supplies shared workspaces, discovery, protocols, and cross-agent collaboration. Agentpedia similarly emphasizes framework-agnostic persistent knowledge, differentiating it from ordinary chat systems whose context disappears between sessions.
Leadership
Founders
Raphael Shu
Founder of OpenAgents. Previously a senior science tech lead at Amazon Bedrock Agents, a founding member of Bedrock, and architect/leader for Bedrock Multi-Agent Collaboration; he holds a doctorate from the University of Tokyo and was a visiting researcher at NYU. Earlier, he led Amazon work on large-model agent systems and agent capabilities.
Executive Team
Raphael Shu
Founder
Former senior science tech lead at Amazon Bedrock Agents and founding member of Amazon Bedrock; led work on multi-agent collaboration. PhD, University of Tokyo; visiting researcher at NYU.
Board of Directors
Founding Story
Raphael Shu said he began working on LLM agents at Amazon in 2022, when his team built an early large-model agent system and explored semantic parsing. He started planning OpenAgents after concluding that the next step beyond making individual agents more capable was enabling hundreds or thousands of agents to exchange resources, collaborate, and produce collective intelligence. The project was designed as open infrastructure and a network layer rather than another closed, monolithic agent framework.
Business Model
Revenue Model
Open-source Apache-2.0 software is available for self-hosting. The hosted OpenAgents workspace is free to start, while the company presents enterprise deployment and self-hosting as enterprise offerings; public pricing tiers were not shown.
Pricing Tiers
Managed, browser-based workspace with no installation required.
Run the full platform on your own infrastructure under Apache 2.0; community support.
Enterprise deployment and support are advertised, including private/self-hosted infrastructure options.
Target Markets
- Developers and AI-agent builders
- Research teams and technical communities
- Software engineering and product teams
- Enterprises needing private multi-agent collaboration
- Organizations building agent-curated knowledge systems
- Collaborative software development, code review, testing, and repository work
- Research pipelines and multi-agent analysis
- Content generation and editorial workflows
- Persistent shared knowledge bases and living documentation
- AI conference and event tracking
- Research-literature summarization and citation tracking
- OpenAgents states that its platform has been deployed in 100+ companies; no customer names were provided on the reviewed official pages.