Seal AI Private
Seal AI Private is Sealmetrics' platform-managed AI provider for LENS AI: your chat questions and AI insights are processed entirely in the EU (Paris), with no prompt retention, and your data never trains third-party models. No API key required — it works out of the box.
It is available as a paid add-on on Growth plans and is included in Scale and Enterprise plans.
What powers it
| Model | gpt-oss-120b — open weights, Apache 2.0 |
| Served by | Scaleway SAS (Iliad group) — Paris, France |
| Retention | None. Token counters only |
OpenAI is not involved and never receives your prompts. gpt-oss-120b is an open-weights model published under Apache 2.0 — a static public artifact loaded onto Scaleway's infrastructure in Paris. There is no call to OpenAI anywhere in the request path, and no contract with OpenAI behind this service.
Scaleway is a named subprocessor in Annex 3 of our DPA, which is the authoritative and always-current list. For why this model was chosen over the alternatives — including the two we tried and replaced first — see Why Seal AI runs on gpt-oss-120b; for how the request path is isolated, see How Seal AI works.
If you use BYOK instead, none of the above applies: your prompts go to your own provider, under your own key and your own contract, and that provider is not a Sealmetrics subprocessor.
LENS AI itself costs nothing extra. It is included in every paid plan, and using it with your own LLM key (BYOK) carries no Sealmetrics charge at all — no add-on fee, no token quota, no packs. You pay only your own provider (OpenAI, Anthropic, Gemini or DeepSeek) for what you use.
The pricing below applies only if you choose Seal AI Private, our managed EU-hosted LLM, instead of supplying your own key. It is a convenience and data-residency option, not a requirement for using LENS.
Pricing
| Contract | Price |
|---|---|
| Annual contracts | €299/month, billed €3,588/year upfront |
| Monthly contracts | €358.80/month |
Activating the add-on mid-cycle is not prorated — billing starts at your next billing cycle. The add-on follows the billing interval of your base subscription.
Availability by plan
| Plan | Seal AI Private |
|---|---|
| Free | Not available |
| Growth | Available as a paid add-on |
| Scale | Included |
| Enterprise | Included |
The token quota
- Seal AI Private includes 5 million tokens per calendar month (UTC). Unused tokens do not roll over.
- Tokens are counted as prompt + completion tokens of each AI request (chat questions and automated LENS insights both draw from the same quota).
- At 80% and 100% consumption, the organization receives an email alert (once per month each).
When the quota runs out
When the monthly quota and any pack balance are exhausted, Seal AI Private stops (hard stop) for both chat and insights, and the product shows a message with a purchase option. Two ways to keep going:
- Buy a token pack (below), or
- Switch to your own key (BYOK) — free of charge from Sealmetrics. The chat's provider selector lets any user fall back to their own OpenAI, Anthropic, Gemini, or DeepSeek key at any time, with no quota. See LLM Providers.
Extra token packs
| Size | 5,000,000 tokens per pack |
| Price | €358.80 per pack (same price on all contracts) |
| Quantity | 1–10 packs per checkout |
| Expiry | Packs never expire — pack tokens are consumed only after the monthly 5M quota is exhausted |
Packs are a one-time purchase (no auto-recharge) made from the Seal AI Usage screen via Stripe checkout. Only organization owners can purchase.
Who can manage it
| Action | Who |
|---|---|
| View the Seal AI Usage screen | Organization owner and admins |
| Activate the add-on / buy packs / cancel | Organization owner only |
| Members | No menu entry; the chat still shows the aggregate quota ring, and on exhaustion members see "contact your admin" |
Cancelling the add-on takes effect at the end of the current billing period.
For developers
Seal AI Private is managed through these API endpoints (JWT with billing:manage — organization owners):
| Endpoint | Description |
|---|---|
POST /billing/seal-ai/subscribe | Add the add-on to the org's subscription (no proration; 403 on Free, 409 if already included in plan or already active) |
POST /billing/seal-ai/unsubscribe | Remove the add-on at period end |
POST /billing/seal-ai/token-pack/checkout | Stripe checkout for 1–10 token packs (entitled orgs only) |
GET /organizations/{slug}/seal-ai/usage | Monthly usage summary — see Organizations API |
When the quota is exhausted, AI endpoints return 403 with error_code: "seal_ai_quota_exceeded" and a structured payload (used_tokens, pack_balance_tokens, effective_quota_tokens, can_buy_pack, entitlement_source, resets_at). See LENS API.
Related documentation
- Seal AI Usage screen — the consumption dashboard
- LLM Providers — Seal AI Private vs. bring-your-own-key
- LENS AI — what the AI can do
- Why Seal AI runs on gpt-oss-120b — the model audit and the trade-offs
- How Seal AI works — the request path and isolation model
- Plans & Pricing — base plan pricing