Plans
The AACsearch plan matrix — Free, Starter, Pro, Business, Enterprise — with limits, feature gates, and the right plan for typical workloads.
AACsearch ships five plans. The matrix below is the canonical reference — the pricing page on aacsearch.com, the dashboard plan card, and these docs are kept in sync from the same configuration in packages/payments/lib/entitlements.ts.
Plan matrix
| Plan | Search units / month | Indexed documents | Indexes | Seats | API keys / index | Connector syncs / month | Analytics retention | Support |
|---|---|---|---|---|---|---|---|---|
| Free | 10,000 | 1,000 | 1 | 3 | 3 | 30 | 7 days | Community |
| Starter | 100,000 | 10,000 | 3 | 10 | 5 | 300 | 30 days | |
| Pro | 1,000,000 | 100,000 | 10 | 25 | 10 | 3,000 | 90 days | Priority email |
| Business | 5,000,000 | 1,000,000 | 50 | 100 | 20 | 30,000 | 365 days | Dedicated |
| Enterprise | Custom | Custom | Custom | Custom | Custom | Custom | Custom | SLA 99.95% |
A search unit is one search request OR one indexed document write. See Usage units for the full unit definitions and how AI usage is metered separately on the wallet.
Feature matrix
| Feature | Free | Starter | Pro | Business | Enterprise |
|---|---|---|---|---|---|
| Public search API | ✓ | ✓ | ✓ | ✓ | ✓ |
| Browser SDK + widget | ✓ | ✓ | ✓ | ✓ | ✓ |
| Filters / sort / pagination | ✓ | ✓ | ✓ | ✓ | ✓ |
| Synonyms | — | ✓ | ✓ | ✓ | ✓ |
| Curations (pin / hide) | — | ✓ | ✓ | ✓ | ✓ |
| Custom origin allow-list | — | — | ✓ | ✓ | ✓ |
| Scoped search tokens (multi-tenant) | — | — | ✓ | ✓ | ✓ |
| Connectors (PrestaShop, Bitrix, …) | — | ✓ | ✓ | ✓ | ✓ |
| Knowledge module (RAG) | — | — | ✓ | ✓ | ✓ |
| AI rerank | — | — | ✓ | ✓ | ✓ |
| Custom branding | — | — | — | ✓ | ✓ |
| Audit log export | — | — | ✓ | ✓ | ✓ |
| SCIM 2.0 | — | — | — | — | ✓ |
| SAML SSO | — | — | — | — | ✓ |
| Data residency (EU/US/RU) | — | — | — | — | ✓ |
| 99.95% SLA | — | — | — | — | ✓ |
Feature gates are enforced at request time by featureGate("synonyms") (and equivalent) middleware in packages/api/modules/entitlements/. The UI also hides controls the active plan does not grant — for example, the Synonyms tab is hidden on Free.
Right-sizing your plan
These are the example shapes we use internally when sizing customers. Treat them as starting points and confirm against your own analytics.
Small catalog (under 1k SKUs, <10 searches/min peak)
Free or Starter. A small e-commerce shop or content site with up to a thousand items easily fits Free for early traffic. Move to Starter when synonyms or connectors are needed.
Growing e-commerce (10k–50k SKUs, 50–500 searches/min peak)
Starter or Pro. Pro adds custom origins, scoped tokens, the Knowledge module, and AI rerank — usually wanted by the time you reach this size. Estimate ~30 searches per visitor per session; a busy storefront does 5–20 million searches/month.
Marketplace / multi-tenant SaaS (any size, multi-org)
Pro or Business. Scoped tokens (Pro+) let you isolate tenants inside a single index. Business is the right floor once you need >10 indexes or >25 seats.
Knowledge base / docs site with AI Q&A
Pro + AI wallet top-ups. The Knowledge module and AI rerank are gated on Pro+, but AI usage is metered separately against the wallet — see Wallet & AI credits. Plan picks the gate; wallet pays for the calls.
Regulated / large enterprise
Enterprise. SSO/SAML, SCIM, data residency, and the SLA only land here. Talk to sales — Enterprise plans are negotiated with custom limits.
Plan change behavior
- Upgrade — takes effect immediately. The new tier's limits and features unlock on the next request after the payment-provider webhook lands. Quotas reset to the new caps; you do not lose the rest of the period's usage budget — it is replaced by the higher cap.
- Downgrade — takes effect at the next renewal so you do not lose features mid-period. If you downgrade below your current usage (e.g. you have 30k documents but the new plan caps at 10k), reads continue but new ingest is blocked until you trim — see Quotas.
- Cancel — keeps the plan active until the end of the paid period. After that you fall to Free with a 7-day grace window for reads (writes stop immediately on cancellation date).
Full timing rules: Upgrade & downgrade.
How AACsearch resolves your plan
Every request goes through:
Auth gate → Feature gate → Quota gate → Rate gate → Search engineThe feature gate looks up your plan via resolveOrgPlan(orgId) (60-second cache, fail-open) and compares it against the matrix. The quota gate checks checkQuota(orgId, "search") against the current period's SearchUsageEvent aggregation.
A full description of the entitlement code paths lives in Plans & Limits (architecture) — that page is the architectural deep dive; this page is the customer-facing plan matrix.
Custom pricing
Enterprise is the only plan with custom limits. For customers on Business who need a one-off lift (e.g. a Black Friday traffic spike), buy overage instead — it's the lighter-weight path.
Related
- Usage units — what counts as a search unit.
- Quotas — soft/hard cap behavior and reset schedule.
- Upgrade & downgrade — plan-change timing and data retention.
- Wallet & AI credits — pay-as-you-go AI usage.
- Plans & Limits (architecture) — entitlement code paths.