Agency tier

iMessage for all your users

iMessage infrastructure for multi-tenant AI products. Provision a sub-tenant per end-user via API, each with its own key, allowlist, and message routing. Built for OpenClaw and Hermes integrators, consumer-AI SaaS, and any agent product where the unit is a single user.

Agency setup uses managed routing through one agency line. Dedicated visible numbers per agent or client are not included by default; bring that requirement to a fit check before onboarding.

Current builder signal

Agency and product builders are checking the first-thread path.

Individual builders affiliated with these teams have started evaluating Claw Messenger. The proof still stays product-level: one controlled message out, then one reply back into the agent path.

  • OpenAI
  • Microsoft
  • Persist Ventures
  • Moveris

Company names describe individual builders' public affiliations, not partnerships, procurement, or endorsements.

Proof before rollout

Prove one client-ready thread first.

Before you package texting for every client, prove the loop once. Send one controlled message from an AI agent, reply from a phone you control, and confirm the reply reaches your agent, webhook, or backend.

First-thread setup

Start with proof, then choose the account path.

After that first thread works, stay self-serve for one small agent, move to agency routing for client or sub-tenant workflows, or talk through a product integration if messaging becomes part of your app.

Good fit for agency

Use the managed path when you build AI agents for clients, can name one first workflow and controlled test number, and expect more than one client, sub-tenant, cohort, or end-user segment.

Better fit for self-serve

Start self-serve when you are one developer or one team proving one agent number. The fastest proof is still API key, test recipient, one outbound message, and one reply.

Start self-serve trial
Not a fit yet

Wait if you only want personal iMessage on Android, need voice calling, need a blast tool, or require every end user to have a visibly distinct dedicated phone number today.

How it works

Three primitives

Parent + sub-tenants

One billing account. Each end-user is a sub-tenant under your parent with its own API key and message routing. You hold the parent key; your end-users never see Claw Messenger.

Programmatic provisioning

Create sub-tenants via API, each with its own key and routing. No manual onboarding per end-user. Webhooks fire when messages arrive and when usage thresholds are hit.

Roll-up billing

Sub-tenant usage rolls up to a single invoice. We bill you; you bill your customers however you want. Resell at any margin.

Pricing

$3 per active sub-tenant

Includes 1,000 messages per sub-tenant per month. Overage at $0.005/message.

Agency

$3/mo per active sub-tenant

Only billed for sub-tenants that send or receive at least one message in the period. Inactive accounts cost nothing.

$250 onboarding fee, or free for founding clients willing to share feedback and try new features early.

Setup includes managed routing through one agency line. It does not include dedicated visible numbers per agent or client by default.

Unlimited sub-tenants

Provision as many end-user sub-tenants as you want. No setup cost per account. Inactive sub-tenants are free.

Managed agency line

One managed line for your agency, routed across your sub-tenants and isolated from self-serve traffic. Per-sub-tenant inbound routing keeps replies pointed at the right agent or customer.

iMessage, RCS, and SMS

One channel, three protocols. Blue bubble where it lands, RCS next, automatic SMS fallback elsewhere. Tapbacks, typing indicators, and read receipts where the recipient supports them.

Managed rollout request

Tell us what comes after your working thread.

If you chose Agency or Product after proving a reply-backed thread, send three rollout details. We'll use them to recommend the right managed setup without asking for message content or private customer data.

This records rollout fit, not message content. Managed setup uses one agency line by default.

Already proved a working thread? Sign in to submit.

Who it's for

Who's building on it

OpenClaw / Hermes integrators

You deploy OpenClaw or Hermes setups for clients. Their next ask is iMessage. Plug in the agency tier instead of building gateway infra. Resell at margin and keep the relationship.

Consumer-AI multi-tenant SaaS

You're building a consumer-AI product where each end-user gets their own assistant. Each one needs private routing into the right agent without seeing Claw Messenger. Programmatic sub-tenant provisioning ships them in seconds, not days.

Education and tutoring platforms

Each student gets a personal AI tutor on iMessage, reachable where they already are. Per-student provisioning, per-cohort rollout, billed only for actively-engaged learners.

Coaching, wellness, accountability

Each user gets their own AI coach in their texts. Daily nudges, check-ins, follow-ups in the same thread their friends are in. They don't leave a thread they live in.

In production

It just works in the background.

Builders run their agents on it every day. Here's what they tell us.

I built an AI agent to help small construction companies manage their office and admin work. The agent's primary mode of communication is text message which makes it natural for the user. I picked Claw Messenger because Anthropic indicated it was the best fit to use with a MyClaw agent. I found it easy to set up and it has been working perfectly.
Dan T.

Ready to talk?

Agency tier is invite-only while we onboard the first wave. Reply with what you're building and we'll get you set up in 48 hours.

Request accessBook a callRead the technical docs
FAQ

Agency questions

Should I use self-serve or the agency tier?
Use self-serve if you are proving one agent number for one team. Use the agency path if the first successful thread needs to turn into client routing, sub-tenant provisioning, or messaging inside your own product. Either way, start with the same proof: one agent message out and one reply back into your agent, webhook, or backend.
What should I bring to first-thread setup?
Bring the agent or workflow you want to test, one phone number you control, the expected reply behavior, and whether the next step is one client, many clients, or an embedded product flow. Do not bring a full production migration as the first step. The first goal is proof, not scale.
Does each end-user, agent, or client get their own dedicated phone number?
No, and we'd rather be straight about it. Sub-tenants route through one managed agency line. Each end-user is isolated by their own API key, allowlist, and routing: when you register an end-user's number, messages between them and the agency line route only to that sub-tenant. Dedicated visible numbers per agent or client are not included by default; if that is a hard requirement, book a fit check before setup.
Can I provision sub-tenants and keys programmatically?
Yes. Create, list, update, suspend, and rotate keys for sub-tenants via the agency API under one parent key. No manual onboarding per end-user. See the technical docs for the endpoints.
Do I need a Mac, iCloud accounts, or BlueBubbles?
No. We manage the iMessage line upstream, so there are no Macs, SIMs, iCloud accounts, or BlueBubbles on your end. Your agents connect to us over an outbound WebSocket.
Where can my agents run?
Anywhere that can hold an outbound WebSocket: a VPS, a Mac mini, a Docker host, or a Raspberry Pi. No public URL or inbound webhook endpoint is required on your side.
Can each sub-tenant restrict who is allowed to message it?
Yes. Each sub-tenant registers the end-customer numbers it talks to. Inbound from those numbers routes to that sub-tenant; unregistered numbers are dropped.
What does it cost?
$3/mo per active sub-tenant, including 1,000 messages each; overage is $0.005/message. You're only billed for sub-tenants that send or receive at least one message in the period. There's a $250 onboarding fee, waived for design partners willing to share feedback and try new features early.
Which features are supported?
iMessage with automatic RCS/SMS fallback, group chats, attachments, tapbacks, typing indicators, read receipts, and delivery-status webhooks.
What happens if a number is throttled or flagged?
We manage line health for you: per-sender caps, gradual warm-up, and automatic throttling of abnormal fan-out. We are not a blast tool, and that is deliberate, because it's what keeps the line trusted and deliverable. If usage hits a sub-tenant's cap, sends are rejected until the next cycle or an upgrade; inbound still reaches you.
Is embedding or reselling Claw Messenger in my product allowed?
Yes. That's exactly what the agency tier is for: your end-users never see Claw Messenger. You bill them however you like.
What data is retained?
We log message content and metadata to route messages and debug issues. If you have specific privacy or data-handling requirements, tell us and we'll walk you through exactly what's retained and for how long.