openclawMU
Install Star on GitHub

Bot developers

Graduate a prototype to a product

You built a single-user assistant that works — now customers want their own, isolated and billed.

Upstream compatibilityPer-tenant tokensQuotasOpenAI-compatible API

The problem

  • Your single-user OpenClaw bot is a hit, and now people want their own instance — but you don't want to maintain a fork per customer.
  • You'd rather not rewrite the agent logic you already have working just to make it multi-tenant.
  • You need paid tiers, but you have no metering and no way to keep one customer's usage from spilling into another's.

With OpenClawMU

  • OpenClawMU keeps every upstream OpenClaw channel and agent feature, then layers tenant token auth and per-tenant isolation on top — your agent behaviour is unchanged.
  • Fork commits are labelled OPENCLAWMU ADDITION so you can rebase against upstream cleanly and keep pulling upstream improvements.
  • The OpenAI- and OpenResponses-compatible tenant API means your existing client code points at the gateway with a token and just works — sessions, memory and quotas apply per tenant automatically.

How it works

Getting there, step by step.

  1. 1

    Move to the MU gateway

    Run your existing OpenClaw configuration on OpenClawMU — the multi-tenant fork keeps the same channels and agent features.

  2. 2

    Turn each customer into a tenant

    Create a tenant per customer; each gets an isolated directory and a token.

  3. 3

    Point clients at the tenant API

    Swap the base URL to your gateway's OpenAI-compatible endpoint and pass the tenant token — no SDK changes for your users.

  4. 4

    Add plans with quotas

    Attach quotas per tenant to back your free and paid tiers, and export usage for billing.

the shape of it 
# one gateway, many tenants
npm install -g openclaw@latest
openclaw onboard --install-daemon

# a tenant per customer / team
openclaw tenants create acme

Frequently asked

FAQ — graduate a prototype to a product

Every question here is also emitted as FAQPage JSON-LD — so Google AI Overviews, Perplexity, and ChatGPT browsing extract the answers cleanly. Crawler-friendly fallback below for screen readers and bots that don't run JS.

No. OpenClawMU preserves upstream OpenClaw's channels and agent features; the fork adds isolation and multi-tenancy around them. Additions are labelled so you can rebase against upstream cleanly.

Related use cases

EXFOLIATE!

Run your own gateway today.

Apache-2.0, self-hosted, no SaaS layer between you and your users. Install the CLI, create your first tenant, mint a token — you're routing traffic in 60 seconds.

Install in 60 seconds Read the source ↗