Our take
Activeregular commits and releases- Activity
- 82
- Community
- 69
- Issues
- 53
- Pull requests
- 53
+255 stars in the last 30 days
- Merges more pull requests than 91% of the projects we track
- Releases more often than 81% of the projects we track
- Slower to merge pull requests than 83% of the projects we track
AgentTeams is an actively developed option for teams that value visible, human-steerable multi-agent coordination and are prepared to operate its Matrix, gateway, storage, and agent components.
Good fit if
- You want agent work coordinated in Matrix rooms with human oversight built in.
- You need multiple agent runtimes or Kubernetes-native management of Workers and Teams.
- You value centralized credential handling and can operate a local Docker or Kubernetes deployment.
Look elsewhere if
- You need a single agent runtime rather than a platform that orchestrates agent containers.
- Your workflow does not need Matrix-based collaboration or human intervention in agent rooms.
- You cannot support the deployment components required by the local or Kubernetes setup.
All health signals
| Last commit | 2026-10-30 (today) |
|---|---|
| Commits, last 90 days | 100+ |
| Releases, last 12 months | 33 (latest v1.2.5, 2026-10-16) |
| Contributors | 52 (top contributor: 48% of commits) |
| Issues closed, last 90 days | 8+ (typically closed in 8 days) |
| Pull requests merged, last 90 days | 87 (typically merged in 3 days) |
| Project age | 8 months |
Checked on 2026-10-31 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
5,999Forks
740Change since Nov 12, 2026: +6 stars·0 forks
Overview
AgentTeams is a collaborative multi-agent runtime platform built around Matrix rooms. A Manager coordinates Worker agents in shared rooms where people can see their exchanges and step in, addressing the lack of visibility and control that can come with agent-to-agent workflows. It orchestrates agent containers rather than implementing agent logic itself.
The project supports multiple Worker runtimes, including OpenClaw, QwenPaw, and Hermes. It can run locally with Docker or as a Kubernetes-native deployment using a controller and Helm chart, making it relevant to both individual experiments and shared enterprise environments.
Key Features
- Manager-Workers architecture for creating and assigning work to agents.
- Matrix rooms keep human and agent conversations visible and open to intervention.
- Co-located Workers can use different runtimes, including OpenClaw, QwenPaw, and Hermes.
- Higress AI Gateway centralizes LLM traffic and credentials, so Workers use consumer tokens rather than holding real credentials.
- MinIO provides shared file storage for exchanging information between agents.
- Kubernetes controller manages declarative Worker, Team, and Manager resources.
- Element Web and a bundled Matrix server provide a browser-based chat entry point.
Use Cases
- Software teams can assign coding tasks to Workers and review progress in Matrix rooms, with human direction available during execution.
- Teams with varied agent workloads can assign different roles to OpenClaw, QwenPaw, or Hermes Workers while keeping collaboration in one room.
- Organizations with credential concerns can route model and service traffic through Higress rather than distributing real API credentials to Workers.
- Platform teams can deploy and manage agent teams on Kubernetes using the controller and Helm chart.
What you need
Detected in the repository
- Go (main language on GitHub)
- A Helm chart for Kubernetes
- A test suite and automated checks on GitHub Actions
License in plain words
Apache-2.0permissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the license, the NOTICE file and a note of your changes
- Share your changes: no
- Includes an explicit patent grant from the contributors:
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
For a local setup, the README requires Docker Desktop on Windows or macOS, or Docker Engine on Linux:
bash <(curl -sSL https://raw.githubusercontent.com/agentscope-ai/AgentTeams/main/install/agentteams-install.sh)
Then open http://127.0.0.1:18088. For Kubernetes installation, configuration, and Windows instructions, see the README.
Alternatives
- multica: Multica assigns coding issues to agent CLIs in a human team workspace, rather than coordinating Manager and Worker agents in Matrix rooms.
- claude-code-by-agents: claude-code-by-agents coordinates coding agents in a native macOS app with threaded rooms, rather than providing a Matrix-based runtime platform.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| AgentTeams | Go | Apache-2.0 | 5.9k | Active |
| multica | Go | – | 53.2k | Active |
| claude-code-by-agents | Swift | MIT | 892 | Quiet |
Considerations
- This is a multi-component platform, not a standalone agent. Local deployment uses Docker, while shared deployments use Kubernetes and Helm; teams should account for operating the Matrix server, gateway, storage, and controller.
- The project supports several runtimes, but the README identifies DeepSeek Harness as experimental and notes runtime-specific configuration constraints.
- The measured health data shows recent commits and frequent releases, with contributions from dozens of contributors and no strong dependence on a single contributor. The repository also has a substantial open-issue backlog, so adopters should review issue and release activity for their needs.
- The Apache-2.0 license permits broad use, subject to its license terms and notices.
Found this useful?
Share it with someone who would like AgentTeams.