Billing & quotas
How trials, per-seat billing, and the AI generation allowance work.
Trials
Every new organization starts on a 14-day Pro trial with full access to Pro features — no payment method required. When the trial ends, the organization moves to the Free plan unless you upgrade.
Trial extensions. The org owner can extend a trial up to two times, to a 30-day total cap. Extensions are owner-only and recorded in the audit trail.
Per-seat billing
Pro and Business are billed per seat, per month. Your Stripe subscription quantity equals your billable seats:
- A billable seat is an active, non-viewer user.
- Viewer-only accounts don't consume a seat.
- The count is floored at 1.
Seat count stays in sync automatically — inviting, deactivating, or changing a user's role adjusts the quantity, and your next invoice reflects it. Switching plans (Pro ↔ Business) reconciles at the billing provider.
Yearly billing is available and saves roughly 15% versus monthly ($33/seat vs $39 for Pro; $59/seat vs $69 for Business).
Upgrading
Upgrade from Billing settings in the app. You choose the plan (Pro or Business) and interval (monthly or yearly) and complete checkout through Stripe. On success your organization is unlocked immediately — any documents or users locked under Free limits are restored.
Downgrading & cancellation
If a trial expires, a subscription is cancelled, or you downgrade to Free, the Free-tier limits are re-applied:
- Documents beyond the limit become read-only (locked, never deleted).
- Users beyond the single seat are deactivated oldest-first — the org owner always keeps the surviving seat.
- Re-subscribing restores everything; nothing is destroyed by a downgrade.
Organizations using SSO that downgrade below Business are migrated to password login so no one is stranded — see SSO downgrade behavior.
AI generation allowance
Token-consuming AI actions draw from a monthly pool:
Monthly pool = per-seat allowance × billable seats
| Plan | Per-seat allowance |
|---|---|
| Free | 10 / month (single seat) |
| Pro | 50 / seat / month |
| Business | 150 / seat / month |
| Enterprise | Unlimited |
The pool resets on the 1st of each month (UTC).
What counts
Only generations that actually run a model count against the pool:
- ✅ Scope generation
- ✅ Executive summary generation
- ✅ Epic generation
These do not count:
- ❌ Gap detection — deterministic, no model call
- ❌ Reading a cached executive summary (only a fresh generation counts)
- ❌ A generation that was already in flight
When you hit the cap
Over the cap, a generation request is declined (HTTP 429) with a GENERATION_QUOTA_EXCEEDED code and no model runs — you're never billed for tokens past your allowance. Usage resets on the 1st, or you can raise the pool by upgrading or adding seats. Current usage is shown in Billing settings.
Fair-use & abuse prevention
New signups are protected against abuse with device fingerprinting, per-IP signup rate limiting, and disposable-email blocking. These affect only automated abuse, not normal use.