https://github.com/hammermei/agent-chat-gateway
A gateway that bridges chat platforms (Rocket.Chat and more) to AI agent backends (Claude Code, OpenCode)
https://github.com/hammermei/agent-chat-gateway
ai-agent automation chatbot claude-code gateway llm opencode python rocketchat rocketchat-bot self-ho
Last synced: about 1 month ago
JSON representation
A gateway that bridges chat platforms (Rocket.Chat and more) to AI agent backends (Claude Code, OpenCode)
- Host: GitHub
- URL: https://github.com/hammermei/agent-chat-gateway
- Owner: HammerMei
- License: mit
- Created: 2026-03-30T19:44:18.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-04T02:36:46.000Z (about 2 months ago)
- Last Synced: 2026-04-04T23:18:38.627Z (about 2 months ago)
- Topics: ai-agent, automation, chatbot, claude-code, gateway, llm, opencode, python, rocketchat, rocketchat-bot, self-ho
- Language: Python
- Homepage:
- Size: 837 KB
- Stars: 2
- Watchers: 0
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Security: SECURITY.md
- Support: docs/supported-features.md
Awesome Lists containing this project
README
# agent-chat-gateway
[](https://github.com/HammerMei/agent-chat-gateway/actions/workflows/ci.yml)
[](https://github.com/HammerMei/agent-chat-gateway/pkgs/container/agent-chat-gateway)
**Turn your AI agent into a team-shared chatbot — in minutes.**
Already running Claude Code or OpenCode? `agent-chat-gateway` connects it to your team's chat system (Rocket.Chat, and more) so everyone can talk to it directly from chat — no terminal required, no code changes to your agent.
Inspired by [OpenClaw](https://github.com/openclaw/openclaw)'s vision of making AI agents accessible from any messaging app — built for the team layer.
> **How it compares to Claude Code Channels:** Claude Code's native [Channels](https://code.claude.com/docs/en/channels) feature connects a single session to Telegram, Discord, or iMessage — great for personal use. `agent-chat-gateway` is built for teams: multiple agents, multiple chat systems, per-user roles, human oversight for sensitive operations, and shared sessions across your whole workspace.
---
## Features
- 💬 **[Works with Rocket.Chat](docs/supported-features.md#chat-platform-connectors)** — and extensible to Slack, Discord, or any other chat system you use
- 🤖 **[Bring your own agent](docs/supported-features.md#agent-backends)** — Claude Code and OpenCode work out of the box; plug in any other agent too
- 👥 **[User-aware in chat](docs/user-guide.md#user-aware-responses)** — the agent knows who sent each message and can personalize tone, language, and style per person using room profiles
- 🔒 **[Owner & Guest roles](docs/permission-reference.md#roles)** — control who can do what, with different permissions per role
- 🛡️ **[Human oversight for sensitive actions](docs/permission-reference.md#approval-workflow)** — the agent pauses and asks for your approval before executing risky operations like file writes or shell commands
- 🔗 **[Continue your session remotely](docs/user-guide.md#use-case-3--continue-an-existing-agent-session-remotely)** — pin a chat room to an existing agent session and pick up right where you left off, from anywhere
- 📎 **[File attachments](docs/user-guide.md#attachment-handling)** — send files in chat and the agent can read and work with them
- 🧠 **[Context injection](docs/user-guide.md#context-files)** — pre-load domain knowledge, system prompts, or project context into the agent at startup
- ⚡ **[Multiple chat systems at once](docs/user-guide.md#multi-connector-setup)** — connect to several chat platforms simultaneously
- ⏰ **[Built-in task scheduler](docs/scheduling.md)** — let the agent schedule recurring or one-shot tasks directly from chat ("remind me in 5 minutes", "run daily standup at 09:00") without any infrastructure setup
- 🤝 **[Agent-to-agent collaboration](docs/agent-chain.md)** — let multiple AI agents collaborate in a shared room; built-in loop protection keeps conversations bounded and human-observable
---
## What's Supported
| | Supported today | Can be extended |
|--|--|--|
| **Chat platforms** | Rocket.Chat | Slack, Discord, and others |
| **Agent backends** | Claude Code, OpenCode | Any agent with a CLI interface |
---
## Quick Start
### Option A — AI-guided install (recommended)
The easiest way to install is to ask your AI agent to do it for you — it handles dependencies, configuration, and any troubleshooting automatically.
In Claude Code or OpenCode, run this prompt:
```
Please install agent-chat-gateway by following the instructions at https://raw.githubusercontent.com/HammerMei/agent-chat-gateway/main/docs/install-agent.md
```
### Option B — Docker (no local dependencies)
If you'd rather skip installing Python, Node.js, or Claude Code locally, run ACG in a container:
```bash
# 1. Copy the example directory to your deployment location
cp -r docker/docker-compose.example my-acg
cd my-acg
# 2. Fill in your credentials and settings
# .env — Claude Code OAuth token (see file for instructions)
# config/.env — Rocket.Chat URL, username, password
# config/config.yaml — owners, rooms, agents
# 3. Start
docker compose up -d
# Logs
docker compose logs -f
```
See [`docker/docker-compose.example/`](docker/docker-compose.example/) for the full annotated setup — all files are pre-structured and ready to fill in.
> Prefer a native install? See [INSTALL.md](INSTALL.md) for step-by-step instructions.
---
## Running the Gateway
```bash
# Start the gateway
agent-chat-gateway start
# Check status
agent-chat-gateway status
# Stop the gateway
agent-chat-gateway stop
```
See [docs/user-guide.md](docs/user-guide.md) for the full CLI reference, configuration options, and usage examples.
---
## Documentation
| Document | Description |
|---|---|
| [INSTALL.md](INSTALL.md) | Manual installation guide |
| [docs/user-guide.md](docs/user-guide.md) | Configuration reference, CLI usage, and operational guide |
| [docs/architecture.md](docs/architecture.md) | System architecture and module breakdown |
| [docs/permission-reference.md](docs/permission-reference.md) | Roles, permissions, and human oversight deep dive |
| [docs/supported-features.md](docs/supported-features.md) | Supported features, known limitations, and roadmap |
| [docs/requirements.md](docs/requirements.md) | Functional specification and behavioral requirements |
| [docs/scheduling.md](docs/scheduling.md) | Built-in task scheduler — recurring and one-shot jobs from chat |
| [docs/agent-chain.md](docs/agent-chain.md) | Agent-to-agent collaboration — enabling multiple heterogeneous AI agents to coordinate via chat |