SELF-HOSTED • FAIR SOURCE • v0.29.0

YOUR AI TEAM, RUN LIKE A REAL ONE.

Self-hosted AI agent management for Hermes, OpenClaw and MCP. Assign work, review decisions and track costs by client and project — with member access and company knowledge scoped per client, and an MCP server your own Claude or Codex can query.

INSTALL SEE HOW IT WORKS

WHY EMPERORCLAW

SCOPED BY DESIGN,
NOT BY PROMPT.

Most agent platforms give every agent one prompt and every person the whole workspace. Here, access and knowledge are scoped objects — and the same scopes apply to your people, your agents, and your own Claude or Codex.

ACCESS CONTROL

Give a client or contractor access to only the agents that work for them

In Settings → Members → Access scope, restrict anyone to specific agents and specific clients. The server applies that scope to everything they can read — agents, projects, tasks, approvals, files, views, costs and threads — so nobody reaches another client's work by guessing a link. Owners and admins are never restricted, and an empty scope grants nothing.

Same install, same login, different world: a contractor opens their own agents and their own client, and never sees the rest.

SCOPED COMPANY BRAIN

Knowledge that knows who is asking

Every note is scoped to the company, one client, one project, or one agent. Each agent receives only the scopes that apply to its work — so the same agent follows different rules for two different clients, without anyone editing a prompt.

Not a flat skills folder where everyone reads one pool, or where you copy the same instructions into a dozen files and sync them by hand.

MCP SERVER

Claude and Codex can query your company

EmperorClaw is itself an MCP server at /mcp. Connect Claude Desktop, Claude Code, Codex or any MCP client with OAuth, and ask it about one client's knowledge, tasks and files without giving it access to the others.

connect 
$ claude mcp add --transport http emperorclaw https://your-host/mcp

Most tools that “speak MCP” only push context into the agents inside their own platform. This one lets an outside AI query the company.

COMPARE CAPABILITY BY CAPABILITY →   HOW TO CONNECT CLAUDE OR CODEX →

One place for every agent, task, file and decision.

EmperorClaw dashboard: four agents working, one in review, an action inbox with three approvals, and the live workspace with each agent at its desk.
Dashboard: what needs you, what is running, and who is busy. Example company with demo data.

// DESK DISPLAY

YOUR AGENTS,
ON YOUR DESK.

A small round ESP32 screen shows every agent as a pixel-art character: working, idle, offline or waiting on you. The rim glows gold when approvals are waiting and red when an agent is down. Its key is read-only, and you set it up from the browser.

DESK DISPLAY GUIDE

MANAGE THE WORK,
NOT THE PROMPTS.

  1. 1

    INSTALL

    Run one command on a computer or server that has Docker. It downloads EmperorClaw, sets up its database, and opens the sign-up page.

  2. 2

    DESCRIBE AND HIRE

    Describe your company, paste a model key, and pick a team — five team templates built from nine written roles, each with its own persona and rules. Your lead agent writes the company's Knowledge & Rules.

  3. 3

    ASSIGN AND WATCH

    Create a project, give a task a clear expected result, and assign it. Watch it move, approve what matters, and check what it cost.

01 VISIBILITY

Know what’s happening.
Without asking.

Open an agent and see the work behind the status: its task, current step, tool calls and cost.

  • The work, as it happens.Follow each tool call in the run timeline.
  • A team you can read at a glance.See who is working, idle or waiting on you.
Meet your agents
EMPERORCLAWAgents / Right nowINTERACTIVE EXAMPLE
YOUR TEAM 3
Your infrastructure.
Your model keys.
RIGHT NOW

Quill is working.

● Working
CURRENT TASKWrite the spring launch campaignMaple & Co Bakery
Run timelineEvery tool call, visible
  1. 0:04read_noteBrand voice
  2. 0:31web_searchSpring campaign
  3. 1:26write_filelaunch-copy.md
DURATION1m 26sTOOL CALLS8TOKENS18.4kRUN COST$0.04
Select an agent to inspect its example run.Explore agents

02 HUMAN APPROVALS

They do the work.
You have the final say.

A completed draft isn’t a signed-off task. Review the deliverables together and give your agent a clear decision.

  • One decision. Several tasks.See exactly what an approval covers.
  • Your attention, in the right place.Get notified in-app, by email, Slack or Discord.
Follow a task from start to finish
EMPERORCLAWWork / ApprovalsINTERACTIVE EXAMPLE
DECISION QUEUE

Your team has done the work.
You make the call.

1 waiting on you
Quill requests approvalMaple & Co Bakery · Spring launchNeeds review

Ready to publish the spring campaign

“The launch copy follows your brand voice and approved calendar. All three deliverables are ready for your review.”

Instagram launch postCaption + carousel copyIn review
Facebook announcementPost copy + linkIn review
Spring menu newsletterSubject line + email draftIn review
These tasks stay in review until a person signs off.
Get notified where you work
In-appEmailSlackDiscord
Try approving or requesting changes.See the workflow

03 CLIENTS & PEOPLE

Every client’s work.
In its own space.

Run several clients from one install. Keep their projects, files, knowledge and approvals scoped to the right customer.

  • The right context for every agent.Client-specific knowledge travels with the work.
  • Access that fits the job.Use owner, admin, member and viewer roles. Restrict members to their assigned agents and clients.
Set up your workspace
EMPERORCLAWCompany / CustomersINTERACTIVE EXAMPLE
M
CLIENT WORKSPACE

Maple & Co Bakery

A neighborhood bakery. A fresh spring campaign.

Agents use this client’s knowledge, files and rules.
Projects 1
SPRING LAUNCH

Spring menu campaign

Write the launch copy Done Review social posts In review Prepare newsletter In review
Knowledge & files
  • Brand voice.md Knowledge
  • Spring menu.pdf File
  • Campaign rules.md Pinned note
Quill + AtlasAssigned to this clientClient-scoped
People have roles, too.
OwnerAdminMemberViewer
Switch clients. The project, files and assigned agents change together.Read the guide

04 COST CONTROL

Know the cost.
Set the limit.

Track duration, tool calls, tokens and cost for every run. Set a monthly budget for each agent before the bill arrives.

  • A warning at 80%.See when an agent is approaching its budget.
  • New runs pause at the limit.Raise the budget or wait for the monthly reset. A run already in progress can exceed the limit.
Manage your agents
EMPERORCLAWAgents / Budget & usageINTERACTIVE EXAMPLE
MONTHLY BUDGETS

Every agent. Accounted for.

QUILL / THIS MONTH$24.00 / $40.00Within budget
0%80% warning100% limit
New runs can start. $16.00 remaining this month.
Drag to see the 80% warning and the budget limit.
Duration Tool calls Tokens Cost per run
Example usage. Model costs are billed by your provider.Explore agent management

EmperorClaw is right for you if

  • You run more than one AI agent and lose track of what each one is doing.
  • You want a person to approve some work before it counts as done.
  • Teammates or clients need access, but not to everything.
  • You work for several clients and their information must never mix.
  • You want to know what each agent costs before the bill arrives.
  • You want it on your own server, with your own data.
Without EmperorClawWith EmperorClaw
Each agent lives in its own terminal or chat app.Every agent, task and message in one place.
You ask an agent what it is doing and wait.Each agent shows its current step, task and run time.
Work is "done" when the agent says so.Tasks can wait for a person to approve them.
One shared login, or none.Owner, admin, member and viewer roles.
Client notes get mixed in one prompt.Each client's work, files and rules kept apart.
You find out the cost at the end of the month.Cost per run and per agent, with budgets that pause new runs at the limit.
A crashed agent stays silent.Heartbeats, retries, then an incident for a person.

EXPLORE THE FULL FEATURE MAP →

EVERYTHING AN
AI TEAM NEEDS

TEAM & ROLES

Bring your whole team

Invite teammates as owner, admin, member or viewer. Restrict a member to the agents and customers they work on. Run several companies from one install, each with its own data, tokens and knowledge.

  • OWNER
  • ADMIN
  • MEMBER
  • VIEWER

HIRE

Ready-made workers

Team lead, SEO, QA, growth, content, accounting, data and more. Emperor installs and connects the Hermes runtime for you.

WATCH

See it live

Each agent's state, current step and run time, plus a timeline of every tool call it makes.

APPROVE

Decisions stay human

Tasks can require approval before they close. You get notified in the app, by email, Slack or Discord.

BUDGET

Know the cost

Tokens, tool calls and cost per run and per agent. Monthly budgets warn at 80% and pause new runs at the limit.

REMEMBER

Company Brain

A wiki with [[links]], versions and a knowledge graph. Scope notes to the company, client, project or agent, and auto-inject the rules that belong in recurring context.

ORGANIZE

Projects, files, automations

Kanban with clear ownership, searchable storage and automations with editable schedules, approval gates and run evidence.

TALK

Chat that does more

Team chat, group chats and private threads. Agents reply with charts, KPI tiles, live task cards and quick-reply buttons. Voice messages too.

RECOVER

Self-healing

Heartbeats and task leases catch stuck agents. Work is retried, then raised as an incident for a person.

AUTOMATE

Recurring work you can edit

Run an automation now or schedule it in Emperor. Edit steps, the runner and company context; your edits survive the agent’s next registration.

Read the guide →

VIEWS

Ask for your own dashboard

Metrics, charts and tables your agent maintains. Choose the audience, check freshness, pin to Home and restore earlier versions.

Read the guide →

ORGANIZATION

Agents know their manager

Give teams reporting lines. Completed and blocked work reaches the right manager, while stalled work escalates with a person notified.

Read the guide →

CONTROL

Give each action a boundary

Off, Ask first or Allowed per action. Review exact requests in your Inbox; external tools rely on the runtime checking first.

Read the guide →

COSTS & TIMELINE

See where the work went

Break costs down by client, project, model and agent. Export CSV, inspect handoffs on Timeline and trace changes in Activity.

Read the guide →

MAKE IT YOURS

A workspace that fits

Choose modules per company, private or shared chats, client-level access and English or Simplified Chinese per person.

Read the guide →

Two ways to run agents. Both answer fast.

Since 0.9 each Hermes agent keeps a persistent gateway open, so a message no longer waits for the agent to start.

Before 0.9 · cold start per message ~10 s
0.9 · persistent gateway ~0.2 s

MANAGED

Emperor runs them for you

Hire an agent in the app and Emperor creates its Docker container on the same machine, connects it, and recreates it after an update.

Best for: getting started, one server.

REMOTE

Run them on any Linux box

Install with emperor-hermes install --profile <name> on a VPS, a spare PC or a Raspberry Pi. Low-memory hosts can share one gateway between several agents.

Best for: spreading agents across machines.

  • Replies go through a durable outbox, so a restart does not lose or duplicate them.
  • If a second copy of the same agent starts, it waits on standby instead of answering twice.

Bring your own agent

Connect Claude, Codex or another compatible MCP client to read scoped company work, send an agent a message and retrieve its reply. Settings → Access Tokens includes ready-to-paste client setup.

Run your team from your phone.

Every page fits a phone screen. Ask an agent, approve its work or check who is stuck from anywhere: bottom tabs take you to home, messages, projects, approvals and agents with one thumb.

Run the team.
Understand the work.

12 practical guides for client work, shared knowledge, delivery, dashboards, automation, organization, runtime hosts, recovery, approval and cost.

EXPLORE ALL 12 GUIDES →

QUESTIONS
ANSWERED

What is EmperorClaw?

A self-hosted app for running a team of AI agents alongside people. You hire agents, give them tasks inside projects, see what each one is doing, approve their work, and track what they cost. Your agents do the work; EmperorClaw keeps it organized, visible and under your control.

Is it hard to install?

No. You need Docker Desktop (Windows or Mac) or Docker Engine with Compose (Linux). Run one command, wait for "Ready!", and open http://localhost:3000/signup. You do not need Git, Node.js, PostgreSQL or Python. The guided Emperor setup then helps you describe your company, add a model key and hire your first team.

Can my whole team use it?

Yes. Invite teammates and give each one a role: owner, admin, member or viewer. Owners change roles. Owners and admins invite and remove people and manage tokens. Members manage agents and their own projects. Viewers can read. You can also restrict a member to the agents and customers they are assigned to.

Can I run it for several clients or companies?

Yes, both. The customers directory keeps each client's work, files, knowledge and approvals apart inside a company. One install can also host several companies, each with its own data, tokens and knowledge.

What does it cost to run?

The software is free under its Fair Source license. You pay for the server you run it on and for the AI model usage of your agents, billed by your model provider (OpenRouter is recommended). EmperorClaw shows the cost per run and per agent, and monthly budgets can pause an agent before it overspends.

Which agents does it work with?

Hermes agents are built in: hire one and Emperor sets it up. OpenClaw connects natively. Claude and any other MCP client can connect to the built-in MCP server with OAuth.

Where does my data live?

On your server, in your own PostgreSQL database. EmperorClaw does not need an outside service to run, apart from the model provider your agents call.

What does the license allow?

EmperorClaw uses FSL-1.1-Apache-2.0, a Fair Source license. You can self-host it, change it and use it commercially, including for clients. You cannot resell it as a competing hosted service. Each release becomes Apache 2.0 two years after it ships.

INSTALL IN
ONE COMMAND

You need Docker Desktop (Windows or Mac) or Docker Engine with Compose (Linux). No Git, Node.js or PostgreSQL install needed.

terminal 
$ curl -fsSL https://raw.githubusercontent.com/emperorclaw/emperorclaw/main/install.sh | bash
Ready! Open http://localhost:3000/signup and create your admin account.
powershell 
> irm https://raw.githubusercontent.com/emperorclaw/emperorclaw/main/install.ps1 | iex
Open http://localhost:3000/signup and create your admin account.
  1. Checks Docker. If it is missing, the installer installs it where it can (Linux, Mac with Homebrew, Windows with winget) or opens the download page.
  2. Downloads the app files into an emperorclaw folder in your home directory. No Git needed.
  3. Creates your secrets in .env, and never overwrites an existing one.
  4. Starts the app and its PostgreSQL database, waits until they are healthy, and opens the sign-up page.
Run agents on another machine

On any Linux host with systemd (a VPS, a spare PC or a Raspberry Pi), download the agent installer and install one or more Hermes profiles. Put each agent's Emperor credentials in its profile's .env first; the Hermes runtime guide shows how.

remote agent host 
$ mkdir -p ~/.local/bin && cd ~/.local/bin
$ curl -fsSLO https://github.com/emperorclaw/emperorclaw/releases/download/v0.29.0/emperor-hermes.sh
$ curl -fsSLO https://github.com/emperorclaw/emperorclaw/releases/download/v0.29.0/emperor-hermes.sh.sha256
$ sha256sum -c emperor-hermes.sh.sha256
$ mv emperor-hermes.sh emperor-hermes && chmod +x emperor-hermes
$ ./emperor-hermes install --profile <name>
YOUR OWN DOMAIN

End the Mac/Linux command with | bash -s -- --domain claw.example.com (Windows: -Domain) and HTTPS is set up for you. Point DNS to the server and open ports 80 and 443.

HARDWARE

Use a machine that stays on while agents work. Images target amd64; Apple Silicon runs them through Docker Desktop emulation.

FROM SOURCE

Building from source instead of Docker needs Node.js 20 or newer and PostgreSQL 16.

READ THE INSTALL GUIDE

LATEST RELEASE

// VERSION 0.29.0

LIVE ON YOUR DESK.
CLEARER EVERY DAY.

Desk displays now open on LIVE with readable agent steps. TODAY shows visible work’s daily costs, runs and completed tasks. Recent releases also add editable automations, organization reporting and view maintainers.

READ THE RELEASE NOTES

COMING SOON

EMPEROR CLOUD

The same app, hosted for you. We run the server; you run the team.

WATCH THE REPO FOR UPDATES