Billing and pricing

Credits and billing

Credit accounting lives in the platform ledger. Entitlements, plan tiers, and downgrade rules are enforced before traffic reaches the API.

Credit pricing API reference

Spend boundary

Keep plan limits in your adapter

The platform records how much work ran. Plan tiers, entitlements, and downgrade rules belong in your integration layer.

Gate expensive scan paths before calling the API.
Return a clear error when credits are exhausted.
Log creditsConsumed next to each incident id.

Platform limits

Tiers, per-minute ceilings, and NSFW daily pools

Per-minute ceilings depend on lifetime credits purchased. NSFW requests without custom text rules may draw from a shared daily pool. Check live counters with GET /v1/billing/credits.

Tier ceilings apply before traffic reaches moderation workers.
Overflow NSFW pool usage bills at standard rates.
Correlate tier rejections with incident and credit metadata during support.

Credit behavior

Relating credits to incidents

Operators should be able to explain cost from response metadata, incidents, and policy inputs.

Basic text is cheap; custom rules cost moreText-only events with no custom policy use the standard text path and cost about 0.05 credits. Text with banned words, rules, or policy text uses model-backed moderation and costs more.

Credits are charged on moderation workEach event consumes credits based on the scan path used. Media analysis, OCR, and deeper review flows cost more than lightweight text passes.

Custom blocks and template uploads are billableWorkspace-level custom image blocks and scam templates are stored on platform and debit credits when created.

Tier limits stay in the callerUse the returned credit totals for accounting, but keep free, accelerated, pro, ultra, or other product-tier limits in your own adapter before you call platform.

Every sync response includes creditsConsumedThe response gives you per-request usage so you can reconcile moderation cost with your own product-level budgeting.

Insufficient credits should fail visiblyIf the workspace has no balance, the platform rejects the request rather than silently skipping moderation.

Credit pricing

Baseline costs by moderation path

Values are defaults from the platform catalog. Actual usage is reported per response as creditsConsumed.

Text only

Basic text moderation without custom rules.

1.0base credits

Custom rules

Model-backed policy text, banned words, and caller-defined suggestions.

1.0base credits

Media

Image and video paths include heavier analysis work.

1.0 / 2.0image / video credits

Tiers and limits

Lifetime credits purchased set workspace tier

Tiers raise per-minute request ceilings. NSFW free allowances are fixed per workspace. Standard per-request pricing applies outside the NSFW free window.

Pass options.billing.contentClass: "nsfw" on POST /v1/events for NSFW-only traffic with no custom text rules. GET /v1/billing/credits returns tier, limits, and remaining allowance.

NSFW class requirements

No custom text rules, transcript entries, or suggestion schema in platformPolicy.text

Omni-style path only (no custom model pipeline); otherwise normal credits apply

Within the fixed daily NSFW free buckets the request consumes 0 credits; overflow is billed at standard credit rates

Tier	Min lifetime credits	Requests / minute	NSFW free / day (all tiers)
0	0+	200	100,000 · 5,000 · 1,200
1	5,000+	600	100,000 · 5,000 · 1,200
2	25,000+	2,000	100,000 · 5,000 · 1,200
3	100,000+	6,000	100,000 · 5,000 · 1,200
4	500,000+	20,000	100,000 · 5,000 · 1,200
5	2,000,000+	60,000	100,000 · 5,000 · 1,200

Text units accrue as roughly one unit per 4k characters of input text on NSFW-class requests (minimum one unit when text is present).

Detailed table

Credits per moderation request

Primary moderation paths exposed through the public API.

Type	Base	Per unit	Description
Text only (basic)	1.0	Whole-request settlement	Lightweight text moderation with no custom rules.
Text with custom rules	1.0	+1.0 / call	Model-backed policy text, banned words, and custom suggestion output.
Image	1.0	—	Image moderation, OCR, and scam lookup.
Video	2.0	—	Frame extraction and video moderation.
Other file	1.0	—	Non-image and non-video file inputs.

Usage notes

Budgeting

Custom rules, OCR, and media increase cost relative to basic text.

Example

A message with 500 characters of text only costs 1.0 credits. A message with one image and 2,000 characters of text costs 2.0 credits. Internal weights are combined first, then the wallet debit and creditsConsumed are rounded up to the same whole-credit amount.

Custom rule path

Triggered by metadata.platformPolicy.text.rules, transcriptEntries, and suggestionSchema when applicable.