Copied ✓
OPEN SOURCE  ·  SELF-HOSTED  ·  30+ AGENTS INCLUDED

Your AI team.
Ready to work.
One command.

30+ agents pre-installed. Every channel wired.
Your team coordinates itself — you just delegate.

MIT Licensed · Self-hosted · No per-seat fees · Your data stays yours

Home
Team
Capabilities
Channels
Good morning, Felix
4 agents live
4
Active
142K
Tokens
30+
Agents ⚡
Felix · Scout · Counsel · Scribe
all active
30+ Agents pre-installed
40+ Tools available
5 Channels supported
1 Command to deploy

Pre-installed agents — ready the moment you deploy

🔍
Researcherresearch
✉️
Email Assistantinbox
✍️
Doc Writerwriting
💻
Coderengineering
📊
Analystdata
📱
Social Mediamarketing
📅
Plannerautomation
🌍
Translatorlanguage
🎓
Tutoreducation
⚖️
Legal Assistantlegal
💼
Sales Assistantsales
🛡️
Security Auditorsecurity
🧑‍💼
Recruiterhr
🤝
Customer Supportsupport
🧪
Test Engineerqa
🏗️
Architectsystems
🔍
Researcherresearch
✉️
Email Assistantinbox
✍️
Doc Writerwriting
💻
Coderengineering
📊
Analystdata
📱
Social Mediamarketing
📅
Plannerautomation
🌍
Translatorlanguage
🎓
Tutoreducation
⚖️
Legal Assistantlegal
💼
Sales Assistantsales
🛡️
Security Auditorsecurity
🧑‍💼
Recruiterhr
🤝
Customer Supportsupport
🧪
Test Engineerqa
🏗️
Architectsystems

Four services. Pre-wired. You just run it.

We bundled, configured, and connected everything. The complex part is already done.

01
🔵
OpenClaw

Agent gateway & channels. Discord, Telegram, Slack — all wired. Route messages to the right agent automatically.

agent-gateway · channels · routing
02
🟢
OpenFang

Autonomous agent OS. Built in Rust. Blazing fast. Each agent runs sandboxed with full tool access.

rust · autonomous · sandboxed
03
🟣
Paperclip

Workflow orchestration. Org charts, cron jobs, agent pipelines. Model your team as code.

orchestration · cron · pipelines
04
🩷
Hermes

Cognitive layer by Nous Research. Memory, skills, and learning — agents that improve over time.

memory · skills · nous-research

ClawHQ vs. building it yourself.

The four projects exist. Wiring them together — that's the hard part. We did it.

ClawHQ
DIY setup
Agents pre-installed and ready to run
30+ included
Find them yourself
All four services wired and talking to each other
Out of the box
Wire it manually
One command to deploy everything
docker compose up
Hours of setup
Dashboard to manage your whole team
Included
CLI only
Hermes coordinates agents automatically
Pre-wired
Route tasks yourself
HTTPS + TLS certificates, zero config
Caddy auto-renews
Set up yourself
Health checks, auto-restart, monitoring
Configured
Build your own ops

Show up. Delegate. Done.

Every agent knows its job. Hermes routes the work automatically.

📬
Clear your inbox

Email Assistant triages, drafts replies in your tone, and flags what actually needs you. Overnight backlog gone by morning.

🔍
Research anything

Researcher fetches, cross-references, and synthesizes. Ask in plain English — get a sourced report back in minutes.

✍️
Write without writing

Doc Writer turns bullet points into polished documents, reports, and proposals. Your voice, your structure, zero blank page.

💻
Ship code faster

Coder writes, reviews, and debugs. Test Engineer catches what slips through. Pair-program with agents that never sleep.

📊
Turn data into decisions

Analyst reads your numbers and tells you what they mean. Business questions become structured briefings, not raw spreadsheets.

🤝
Coordinate the whole team

Hermes routes your request to the right agents, collects results, and brings back the answer. You delegate once. They handle the rest.

What builders are saying.

"

We replaced three separate AI integrations with ClawHQ. Hermes coordinates everything — I send one message and the research, writing, and code review all happen in parallel. It's like having a real team on call.

👨‍💻
Alex Chen
CTO · Stealth AI startup
"

The self-hosted aspect was the dealbreaker for us. Our clients' data never touches a third-party server. ClawHQ gives us enterprise-grade AI without the enterprise-grade privacy risk.

👩‍💼
Sarah Mercer
Product Lead · Consulting firm
"

docker compose up and everything just worked. The pre-installed agents saved me days of setup. I had Researcher and Coder running in Discord within the hour. This is how it should always be.

🧑‍🚀
Marcus R.
Founder · Solo operator
🔒 Privacy by design

Your data.
Your server.
Full stop.

Every conversation, every file, every agent memory — stored on your infrastructure. No telemetry. No training on your data. No subscriptions that cut off your access.

Open source means you can read every line. Deploy on any server, any cloud, or completely air-gapped.

🏠
Runs on your hardware
One Docker Compose file. Your VPS, your homelab, your cloud account. Nothing calls home.
🔑
You hold your own keys
API keys live in your .env. Secrets never leave your environment. No SaaS middleman.
📖
Fully open source
MIT license. Audit the code, fork it, extend it. No black boxes, no vendor lock-in, ever.
✈️
Air-gap capable
Run with local models via Ollama. Fully offline operation for sensitive environments.

Everything your agents need.

Multi-Provider LLMs

Every frontier model. One interface.

Switch providers per-agent or per-task. Route cheap tasks to fast models, complex reasoning to the frontier.

Anthropic OpenAI Groq OpenRouter Ollama Gemini Mistral + more

Persistent Memory

Agents that remember.

Vector + episodic memory. Agents recall past conversations, learn preferences, and build context over time.

Workflow Automation

Scheduled. Triggered. Automated.

Cron jobs, event triggers, and multi-step pipelines. Your agents work even while you sleep.

Cross-Platform Channels

Meet your team where you already are.

One config. Every platform your team uses, wired to the same agent backend.

💬 Discord ✈️ Telegram 🟣 Slack 📱 WhatsApp 🔒 Signal

Sandboxed Execution

Safe by default.

Every tool call runs in an isolated sandbox. File system, network, and process boundaries enforced per-agent.

Skills Marketplace

40+ tools. Plug in more.

Web search, code execution, calendar, email, databases — and a simple API to add your own.

From zero to your AI team
in three steps.

No cloud lock-in. No per-seat fees. No complicated setup. Your infrastructure, your data.

01

Deploy

Add your API keys to .env and run one command. All four services start, connect, and health-check themselves. No manual wiring.

$ docker compose up -d
02

Meet your team

Open the dashboard. 30+ agents are already there — researcher, coder, writer, analyst and more. Connect your channels (Discord, Telegram, Slack) in one click.

your agents are waiting
03

Just delegate

Message Hermes — your coordinator — with any task. It routes to the right agent, collects the result, and reports back. You never touch a terminal again.

your 24/7 AI coworkers

Early access

Your AI team
is waiting.

Join the waitlist. Be first to know when ClawHQ is ready to deploy.