The open-source, multi-agent alternative to Claude Tag.
@ any agent. Wherever work happens, your agents work alongside
your team and each other, learning as they go.
** Website** ·
·
Blog Why AgentConnect? · Get started · Community · Contributing · Architecture
AgentConnect is an open-source platform where teams and multiple AI agents work together across Slack, Telegram, Discord, Lark, GitHub, and GitLab. Bring Claude Code, Codex, Grok Build, DeepSeek, Pi, or any ACP-compatible agent into the conversations and workflows your team already has open.
Give each agent a role, then let people and agents collaborate in shared conversations. Agents can call one another and remember what they learn, and work can begin from a message, issue, pull request, webhook, or schedule.
▶
Watch the two-minute introduction
AI agents are getting better at doing work. The harder problem is making multiple agents work well with a team—and with each other.
Most agents still live like personal tools, in one person's terminal. Teammates can't see what an agent is doing, can't take over a session, can't review its output, and the context it builds stays on one laptop. So every team writes the same glue: message channels, cron jobs, credential handling, context stitching.
AgentConnect turns that glue into a platform.
Teams use it to:
Triage issues together. People and agents investigate in one shared thread, bring in the right specialists, and keep the fix and verification visible from start to finish.Support across trusted workspaces. Start a support conversation in Telegram, involve engineering from a trusted Slack workspace, and return the resolution where the conversation began.Run recurring operations. Start work from a schedule or webhook, bring exceptions into a shared conversation, and keep the human decision visible.Keep private forks current. Subscribe to upstream changes through GitHub, a GitHub subscription in Slack, webhooks, or schedules. Let agents assess the impact, prepare and test relevant updates, and bring them to the team for review.Run customized code review. Run a general reviewer on every pull request, then bring in architecture or security reviewers only when a change needs them. Each reviewer can use its own model, instructions, repository access, tools, and sandbox policy.
Work as one team, on any runtime. Create agents with different roles and let them call on one another, while people follow along in the conversations where the work happens. Claude Code, Codex, Grok Build, DeepSeek, Pi, and any other ACP-compatible runtime run side by side, and changing one does not rebuild the workflow around it.Keep work where it happens. Link agents to bots in Slack, Telegram, Discord, and Lark, or to repositories and workflows on GitHub and GitLab.Choose the right agent for every job. Configure each agent's runtime, model, workspace, tools, and machine independently.Carry context forward. Give each agent its own memory and skills, and publish reviewedKnowledgethat every agent can find on demand.Set clear boundaries. Decide who can see each agent and session, which repositories and tools it may use, and which other agents it may call.Stay in control. Self-host the Apache-2.0 stack and keep agent execution and workspaces in the environment you operate.
One place to operate your whole fleet—agents, sessions, schedules, tools, knowledge, and daemons:
Don't just take our word for it. Ask your favorite AI about AgentConnect:
Start the Web console, Control Plane, Relay, and PostgreSQL with Docker Compose:
git clone https://github.com/agentconnect-md/agentconnect.git
cd agentconnect
docker compose up -d --pull always
Open http://localhost:3000
, add a daemon from the console, run its generated
command, and create your first agent. The default stack listens only on
127.0.0.1
and uses local no-auth mode for evaluation.
The official Helm chart is published on every release under the same version number as the release itself:
oci://ghcr.io/agentconnect-md/charts/agentconnect
A cluster install is three steps — create the namespace and its secrets, write the values file for your deployment, then install — so follow the Kubernetes guide rather than a one-liner. The chart source lives in charts/agentconnect.
Setup Server runs in the base stack at http://localhost:8091
, always on
loopback. It configures browser authentication through Logto, the deployment's
GitHub, Slack, Google, and Lark / Feishu apps, the sign-in methods you show, and
preset-agent behavior. Deployment topology — images, ports, public URLs,
database and bootstrap secrets — stays in compose.env
or your Helm values.
Prefer to be walked through it? This repository ships a setup skill at
.claude/skills/agentconnect-setup
, also exposed at .agents/skills
, so Claude Code, Codex, and other agent harnesses pick it up. Open your coding agent in the checkout and ask it to set up AgentConnect: it runs the guide as an interactive tutorial and verifies each checkpoint before continuing, and never asks you to paste secrets into chat.
For authentication, public URLs, Linux sandbox requirements, provider apps, image pinning, secrets, and optional Mem0 configuration, follow the AgentConnect OSS guide.
If AgentConnect looks useful, here's how you can support the project:
- ⭐ to help more teams discover the project.Star this repo - 💬 Join the Slack communityfor questions, ideas, and product discussions. - 📣 Share AgentConnect on X,LinkedIn,Reddit, orLINUX DO. - ✍️ Write a review or tutorial on DEV,Medium, or your blog. - 🙋 Open an issueto report bugs or suggest features. - 📺 Follow updates on XandYouTube.
Contributions are welcome. Development needs Node >= 24.12.0 and pnpm 11;
pnpm install && pnpm dev
runs every package in watch mode. The contributing guide covers the full setup, the pull request conventions, the monorepo layout, and the deeper design documentation.
Thanks to everyone who has already contributed:
| Component | Responsibility |
|---|---|
| Daemon | |
| Runs placed agents over daemon-owned ACP, owns workspaces and session state, maintains direct platform connections and schedules, and sends model-provider traffic directly | |
| Relay (optional) | |
| Accepts callback-based ingress and webchat, proxies centrally managed MCP and OpenConnector access, and forwards message ingress directly to the owning daemon without durable storage | |
| Control Plane + Web UI | |
| Manages authentication, configuration, placement, permissions, metadata, and observability; stores explicitly approved organization knowledge/skill revisions and otherwise proxies bounded daemon reads on demand |
Live platform messages and ACP update streams stay on the daemon/relay data plane. Apart from explicitly approved organization knowledge and bounded skill bundles, the Control Plane stores coordination metadata—not message bodies, attachment bytes, pending Dream proposals, or ACP session streams. If it is temporarily unavailable, established sessions and daemon-local schedules continue; new assignments and configuration changes resume after reconnection.
See the architecture design for the complete message paths, trust boundaries, and failure model.
AgentConnect is available under the Apache License 2.0.