Limits and what they mean

Every locked button in the console is one of the allowances below. This page says what each one governs, so a refusal tells you something instead of just stopping you.

Your own numbers and the price of every plan are on Billing in the console. It shows each allowance as used of total, with what remains, and what moving up would cost. This page doesn’t repeat any of it, because a number copied into documentation is a number that goes out of date.

If you haven’t signed up yet, Billing and Plans compares the Home and Work plans. You can still try a public hub in your browser for nothing, and the Free plan’s allowance is below.

How Limits Behave

Your actual limits are in Billing, as a used/total figure per allowance. At the limit, the action is refused and the message says which allowance is full; nothing that already exists stops working. For more room, change plan in Billing or free something you’re not using. To stop paying, Billing → Manage billing opens the Stripe portal.

What Each Allowance Governs

Allowance	What it counts	You meet it when
Public hubs	Private hub capacity available on your plan.	Creating one more hub than the plan carries.
Runtime capacity	Skill sets that can run private workspace workloads.	Adding a runtime group beyond the plan.
Autoscaling hubs	Hubs allowed to scale to paired workers.	Turning on autoscaling for one hub too many.
Clients	Client identities that can connect to public or workspace hubs.	Pairing one more device, app, or agent.
Access policies	Access-rule records available for workspace access control.	Creating another access rule.
Monthly intents	Questions answered across every client, per month.	Heavy shared use. Resets on a stated date.
Daily intents	The same, per day. A safety rail rather than a budget.	A runaway client or a load test.
Custom domains	Whether managed hub addresses can be mapped.	Not a count — either included or not.
API tokens	Active tokens, counted separately from clients.	Minting one more token than the plan allows.

On The Free Plan

	Free
Connections	1
Active API tokens	1
Intents per day	100
Personal memory	1 item
Public hubs	Browse and connect to any of them
Your own hubs	Not on this plan

One connection means one device at a time. Rather than upgrade to pair a second phone, you can move the connection you already have — the app offers that in a tap, and it doesn’t change the count.

When You Reach One

The refusal names the allowance, the limit, and what you’ve used. It’s an HTTP 403 carrying code, resource, limit, used and plan beside the sentence, so your own code can tell a plan limit apart from a fault:

{
  "detail": "Free plan allows up to 1 connection.",
  "code": "plan_limit",
  "resource": "client",
  "limit": 1,
  "used": 1,
  "plan": "Free"
}

resource is one of client, api_token, hub, runtime_group or acl.

Memory is the one allowance with its own refusal. A second memory item on Free answers HTTP 409 with the code memory_limit, and names Home as the upgrade; see Free Tier.

Note

Reaching a limit never stops what is already running. A hub that’s answering keeps answering; a paired client stays paired. Only the new thing is refused.

Changing Or Ending A Plan

Billing is handled through Stripe, so the plan, the payment method, the invoices and the cancellation all live in one place.

1. Open Billing. It shows the current plan and whether the subscription is active.
2. Choose Manage billing. This opens the Stripe customer portal for your workspace.
3. Change or cancel there. Invoices and payment methods are in the same portal.
4. Come back to Billing to confirm the limits the console shows now match what you chose.

To withdraw from a new contract within 14 days, or to cancel without signing in, use Withdraw or Cancel a Plan; both send an email receipt.

Cancelling ends the subscription. It doesn’t delete your account or your data — Profile and security covers exporting your data and deleting the account itself, which are separate and deliberate acts.

Rate Limits Are Separate

Plan allowances count things. Token rate limits count requests per minute and are covered in API tokens. A 429 is a rate limit; a 403 with code: plan_limit is an allowance.