Limits and plans

Every organization runs under hard caps. This page lists them, tells you what you get back when you hit one, and explains which limits are billing (upgrade) versus blast-radius (wait).

Plan tiers

	Free	Pro	Team
Price / month	$0	$79	$299
Runs / month	120	200	1,000
Identities	9	25	100
Outbound emails / month	150	1,000	5,000
Plan runs / day	30	200	1,000
Effective runs / day	30	50	50

A run is one registration, login, or connect attempt driven by the Sente browser. Failed runs never count against any quota — you are not charged capacity for our failures. A run that is queued, running, blocked, or completed does count. The effective daily number is the lower of your plan’s daily allowance and a platform-wide per-organization brake of 50 runs / 24 h — a rolling window, not a calendar day. The brake exists so a runaway agent can’t burn a month of capacity in one loop. If a legitimate workload needs it raised, email support@sente.run; design-partner pilots are invoiced manually anyway. Monthly counters run against your billing window (a rolling 30-day window if you have no subscription).

Why the caps exist

All identities send and receive on one shared Sente-operated domain. Email deliverability is reputation-based and the reputation is shared: one tenant sending spam gets the domain flagged, and a flagged domain means verification emails stop landing for every customer. The email caps are the blast-radius guard for that shared resource, not a growth lever. The run caps are the same logic applied to third-party apps — abusive registration volume degrades the platform for everyone. See the Acceptable use policy.

Limit error codes

Every cap returns a structured error with a code and the limit that was hit.

Status	Code	Meaning	What to do
`429`	`IDENTITY_LIMIT_EXCEEDED`	Your plan’s identity cap. Also returned when connect/register would have to auto-provision an identity and can’t.	Delete unused identities, pass an existing `identityId`, or upgrade.
`429`	`RUN_LIMIT_EXCEEDED`	The daily or the monthly run cap — `limit` tells you which.	Daily: wait for the rolling window. Monthly: upgrade.
`429`	`SEND_LIMIT_EXCEEDED`	The 100-per-24 h outbound send brake.	Wait for the window to roll, or contact support.
`402`	`PLAN_EMAIL_LIMIT_EXCEEDED`	Your plan’s monthly email cap. `402`, not `429`, because waiting won’t clear it inside the billing window.	Upgrade the plan. The body carries `planId`.

{
  "error": "monthly registrations & logins limit reached (120 per month on your plan)",
  "code": "RUN_LIMIT_EXCEEDED",
  "limit": 120
}

Reading and waiting are free: GET /v1/messages, GET /v1/messages/wait (the OTP / magic-link long-poll), GET /v1/runs/:id, and GET /v1/audit-events consume no run or send quota.

Operational limits

Not billing — these are the timeouts and mutual exclusions the platform enforces while work is in flight. Defaults; all are tunable per deployment.

Limit	Value	What happens at the edge
Blocked-run hold	~10 min	A run waiting for a human fails with `BLOCKED_TIMEOUT`. See Human takeover.
Wait for a verification email	8 min	The run fails with `VERIFICATION_TIMEOUT`.
Browser session per run	20 min	The run fails with `AUTOMATION_STUCK`.
Driver heartbeat	5 min stale	The run is recovered and failed with `AUTOMATION_STUCK` rather than hanging.
Session freshness	10 min since the last confirmed login	`getSession` re-logs in first and returns `409 SESSION_STALE` with a `runId` to poll. See Sessions.
Live sessions per identity	1	A second request gets `409` (`SESSION_ALREADY_OPEN`, or `SESSION_PENDING` / `IDENTITY_BUSY` if a run is in flight).
Concurrent runs per identity	1	Calling register/login/connect while a run is live returns that run instead of starting another — the calls are idempotent, so a retry is safe.
Platform run concurrency	a global ceiling	Excess runs stay `queued` and start as capacity frees. Nothing is dropped; a busy period shows up as queue latency.
Inbox long-poll	1–60 s per call	`timeout` above 60 is a `400`. Loop with the same `since` to wait longer.
Message list page	1–200	Default 50, newest first.

Next steps

Acceptable use

What the caps are protecting.

Security model

Where credentials live and who can read them.

API overview

Auth, base URL, and the shared error shape.

Get started

Accounts

Runs

Email

Events & audit

Trust

Limits and plans

Plan tiers

Why the caps exist

Limit error codes

Operational limits

Next steps

Acceptable use

Security model

API overview

​Plan tiers

​Why the caps exist

​Limit error codes

​Operational limits

​Next steps

Acceptable use

Security model

API overview

Plan tiers

Why the caps exist

Limit error codes

Operational limits

Next steps