TypeScript multi-agent framework that runs in your own environment: consequential actions wait for approval and every run leaves a verifiable record. Describe the goal, not the graph. 13 built-in providers (Claude, OpenAI, Gemini, DeepSeek and more) plus any OpenAI-compatible endpoint, local models included.
Open Multi-Agent is a TypeScript framework that takes a plain-language goal and automatically decomposes it into a directed acyclic graph of tasks at runtime, assigning each node to a named specialist agent, parallelizing independent branches, and synthesizing the final result through a coordinator agent. Developers define a team of agents with per-agent model assignments, system prompts, and explicitly opted-in tools such as `bash`, `file_read`, `file_write`, `file_edit`, and `grep`, while the coordinator builds the execution graph without requiring developers to enumerate nodes and edges by hand. It connects to Claude through the Anthropic API, using models like claude-sonnet-4-6, and also supports OpenAI, Gemini, Grok, DeepSeek, and local models via Ollama. A built-in post-run dashboard replays completed runs with per-node assignee, status, token breakdown, and agent output logs. With only three runtime dependencies, the package drops into any Node.js 18+ backend and suits TypeScript developers who need adaptive multi-agent orchestration without hand-wired workflow definitions.
- ✓Open-source license (MIT)
- ✓Actively maintained (<30d)
- ✓Healthy fork ratio
- ✓Clear description
- ✓Topics declared
git clone https://github.com/open-multi-agent/open-multi-agent && cp open-multi-agent/*.md ~/.claude/agents/Subagents overview
What people ask about open-multi-agent
What is open-multi-agent/open-multi-agent?
+
open-multi-agent/open-multi-agent is subagents for the Claude AI ecosystem. TypeScript multi-agent framework that runs in your own environment: consequential actions wait for approval and every run leaves a verifiable record. Describe the goal, not the graph. 13 built-in providers (Claude, OpenAI, Gemini, DeepSeek and more) plus any OpenAI-compatible endpoint, local models included. It has 6.9k GitHub stars and its last recorded update is dated 2026-09-09.
How do I install open-multi-agent?
+
You can install open-multi-agent by cloning the repository (https://github.com/open-multi-agent/open-multi-agent) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is open-multi-agent/open-multi-agent safe to use?
+
Our security agent has analyzed open-multi-agent/open-multi-agent and assigned a Trust Score of 97/100 (tier: Verified). See the full breakdown of passed checks and flags on this page.
Who maintains open-multi-agent/open-multi-agent?
+
open-multi-agent/open-multi-agent is maintained by open-multi-agent. The last recorded GitHub activity is dated 2026-09-09, with 6 open issues.
Are there alternatives to open-multi-agent?
+
Yes. On ClaudeWave you can browse similar subagents at /categories/agents, sorted by popularity or recent activity.
Deploy open-multi-agent to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/open-multi-agent-open-multi-agent)<a href="https://claudewave.com/repo/open-multi-agent-open-multi-agent"><img src="https://claudewave.com/api/badge/open-multi-agent-open-multi-agent" alt="Featured on ClaudeWave: open-multi-agent/open-multi-agent" width="320" height="64" /></a>More Subagents
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
The agent that grows with you
Java 面试 & 后端通用面试指南，覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
The agent engineering platform.
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.