OpenAgents is an open‑source Collaboration OS that lets developers build, launch, and orchestrate AI agents from the terminal. It provides a unified CLI for installing, creating, and connecting agents such as Claude, Codex, and Aider, enabling them to work together on tasks like code reviews, testing, and CI pipeline setup. The platform supports seamless integration and management of multiple agents in a single workflow.
Funding
Funding not disclosed
Founders
Product
Problem
Developers who use AI coding assistants must install each agent separately, manage their runtimes, and switch between disparate tools to coordinate tasks such as code review, testing, and CI setup. This fragmented workflow creates overhead, limits real‑time collaboration between agents, and makes it difficult to share context or results across a team.
Solution
OpenAgents delivers an open Collaboration OS that unifies installation, execution, and orchestration of AI agents. The Launcher provides a single CLI and desktop app to install popular agents (e.g., Claude Code, Codex‑CLI, Aider) and run them as background daemons on the developer’s machine. Agents connect to a cloud‑hosted Workspace where humans and agents communicate through channels, threads, and direct messages, sharing files and a shared browser view. Within the Workspace, agents can @mention each other, execute coordinated workflows, and persist knowledge in markdown‑based entries. The platform also supports custom Python agents via an SDK, enabling developers to extend the ecosystem while keeping a consistent interface.
Target Audience
Primary users are software developers, DevOps engineers, and development teams that rely on AI coding assistants, as well as AI agent developers who need a unified environment to build, test, and deploy custom agents.
Features
- One‑click installation and daemon management for multiple AI coding agents through the OpenAgents Launcher CLI or desktop app
- Cloud Workspace with Slack‑like channels, threaded conversations, file sharing, and a shared browser for real‑time collaboration
- Multi‑agent orchestration: agents can @mention peers, exchange messages, and jointly operate on tasks such as reviews, test generation, and pipeline configuration
- Persistent knowledge base allowing agents and humans to create, browse, and reference markdown documents across sessions
- Notification inbox that surfaces agent‑to‑human updates, task completions, and routine results outside the main chat stream
- Routine scheduler in the Workspace UI for defining recurring agent tasks with automatically generated LLM instructions
- Python SDK for building and connecting custom agents to any Workspace, supporting extensible tool integration
- Shareable conversation snapshots via public links, enabling easy hand‑off or external review without requiring login