We are currently building something exciting and will be launching soon. Registration is temporarily closed. Launching soon — registration temporarily closed.

PRICING

Pay for the runs. Get the humans included.

Every paid plan includes all 12 GA analyzers, verification receipts, and real human help getting thresholds right the first time - because decisions this important shouldn't be self-serve-only.

What a run buys: a versioned semantic signal under a pinned configuration, with a signed receipt you can replay while the configuration is supported under the published support policy or your contract - and verify offline at any time. Not just a score - evidence.

Request access · No credit card required · 5K analyzer runs each month

Free

$0

Explore the API

  • ✓ 5K analyzer runs/mo
  • ✓ Hard cap at quota?
  • ✓ 2 RPS · 1 API key
  • ✓ All 12 GA analyzers & their supported languages
  • ✓ Verification receipts
  • ✓ Community support + docs
  • ✓ Daily evaluation limits apply?
Request access

Need a deeper evaluation? Request an extended allocation - we grant larger quotas manually for legitimate evaluations.

Build

$149/mo

For teams validating a use case

  • ✓ 50K analyzer runs/mo
  • ✓ $2.50 / 1K runs overage?
  • ✓ 10 RPS · 3 API keys
  • ✓ Standard analytics
  • ✓ Email support - within 1 business day
  • ✓ Threshold-tuning template + one email review
Request access

Growth

$349/mo

For teams past pilot, not yet at scale

  • ✓ 150K analyzer runs/mo
  • ✓ $2.00 / 1K runs overage?
  • ✓ 25 RPS · 8 API keys
  • ✓ Advanced analytics
  • ✓ 30-min onboarding call with a solutions engineer
  • ✓ Quarterly office hours
Request access
Most common at 250K+ runs

Production

$799/mo

For teams shipping to real users

  • ✓ 500K analyzer runs/mo
  • ✓ $1.50 / 1K runs overage?
  • ✓ 50 RPS · 15 API keys
  • ✓ Advanced analytics + request evidence
  • ✓ Named customer engineer
  • ✓ Threshold review at 30/60/90 days
  • ✓ Slack Connect channel
Request access

Scale

$2,999/mo

For high-volume production

  • ✓ 2.5M analyzer runs/mo
  • ✓ $0.90 / 1K runs overage?
  • ✓ 200 RPS · 50 API keys
  • ✓ Full analytics + governance exports
  • ✓ Named CSM + technical account manager
  • ✓ Quarterly business review
  • ✓ Customer-local benchmark evaluation - runs in your environment; only agreed aggregates leave
Request access

Automatic overage on paid plans - no hard quota stop. Configure a spend cap if you need one. Annual billing saves 15%.

Enterprise

Annual agreement

For organizations putting reproducible semantic analysis into regulated or org-wide decision workflows. Enterprise sells governance, capacity, a dedicated implementation team, and an annual configuration audit - not just analyzer runs.

✓ Contractual SLA with service-credit remedies & reserved throughput
✓ Contractual guarantee: pinned configurations stay supported for the life of the contract — subject to the documented security-suspension rule (an unsafe execution environment pauses execution, never silently changes behavior; verification continues)
✓ Governance & audit exports
✓ Dedicated implementation team
✓ Annual configuration audit
✓ SSO/SAML, countersigned DPA & security review
✓ Deployment architecture review under NDA (dedicated or customer-cloud modes are scoped case-by-case — not a standard feature)
✓ Custom taxonomy engagement - separately scoped SOW; the data boundary is agreed in writing before any work begins

The standard Data Processing Agreement (v2.0) applies to every plan - Enterprise adds countersignature and security review.

Pricing

By agreement

Contact Sales

Consulting engagements have explicit data boundaries

Evaluation, policy tuning, and detector development are different activities with different data rules. "Custom" work never quietly turns into data collection - the boundary below is the contract, and our zero-retention and no-training commitments apply unless a countersigned SOW amends them in writing.

Customer-local evaluation

Benchmarks and diagnostics run against your data inside your environment.

Boundary: Evaluation tooling (eval-toolkit) executes where your data lives. Only agreed aggregates and diagnostic output - counts, scores, confusion summaries - leave that environment. Sentiqra receives no case corpus.

Never: No customer text enters Sentiqra systems, evaluation corpora, or detector development.

Customer-specific policy configuration

Threshold, routing, and review-workload assistance for analyzers you already run.

Boundary: We work from your measured results, receipts, and policy goals - not your case corpus. The detector, taxonomy, and scoring configuration are unchanged.

Never: No detector or taxonomy changes; no customer data collection.

Bespoke detector or taxonomy development

Separately reviewed statement of work. Not included on any plan.

Boundary: No data is accepted before a countersigned SOW specifies permitted inputs, artifact ownership, retention, isolation, configuration lineage, and prohibited reuse.

Never: Customer content stays out of detector development by default - a custom label definition does not require your case history.

Please do not email datasets or paste raw case material into support tickets or contact forms - we never request case collections through those channels. Any bespoke engagement starts with scope, not data.

On every plan, including Free: all 12 GA analyzers · every language each analyzer supports (43 platform-wide, 8–41 per analyzer) · verification receipts · immutable configuration IDs · zero content retention · no training on your data

Performance targets vs. contractual SLA: the published p95 <500ms latency and 99.9% uptime figures on the performance dashboard are targets that apply to every plan - not contractual commitments. Contractual SLAs with service-credit remedies apply to Enterprise agreements only.

Not on any public plan: 1 restricted analyzer - Information Warfare - is deployed under custom agreements only. Request an access briefing.

Learn more about data handling →

WHEN TO MOVE UP

The honest break-even map

The run volume where the next plan's base price beats your current plan's base + overage. Pure cost math - most teams upgrade earlier for the support tier.

If your monthly runs exceed… You'll save money on…
130,000 Growth
375,000 Production
1,967,000 Scale
Multi-million, or regulated deployment Enterprise - Contact Sales

We publish this because surprise overage bills churn more customers than honest upgrade math does.

HOW BILLING WORKS

Simple usage-based billing

1 analyzer run = 1 text analyzed by 1 analyzer

1 run

Analyze one text with Trust & Safety

2 runs

Analyze the same text with Trust & Safety + Cybersecurity (maximum 2 analyzers per request)

  • ✓ One analyzer run covers up to 4,096 Unicode code points of text.
  • ✓ The number of signals returned does not affect billing.
  • ✓ Failed analyzer executions and valid idempotent retries are not billed.

ESTIMATE YOUR COST

Workload calculator

The full comparison version lives at /cost-calculator - this is the same math.

Estimated analyzer runs 700,000
Recommended plan Production
Plan$799/mo
Included runs500,000
Overage200,000 × $1.50/1K

Estimated monthly total

$1,099

Contact Sales Request access

WORKED EXAMPLES

What real workloads cost

Pick the example closest to yours. The math is the same math the calculator runs.

Marketplace moderation

Trust & Safety on listings + messages

250K texts/mo × 1 analyzer250K runs
Growth $349 + 100K overage × $2.00$549/mo

→ Growth - cheapest cover; onboarding call included

Multi-domain safety pipeline

Trust & Safety + Cybersecurity on uploads

350K texts/mo × 2 analyzers700K runs
Production $799 + 200K overage × $1.50$1,099/mo

→ Production - named engineer + 30/60/90 threshold reviews

SOC alert triage

Cybersecurity analyzer on inbound alerts

60K alerts/mo × 1 analyzer60K runs
Build $149 + 10K overage × $2.50$174/mo

→ Build - cheapest cover at this volume; Growth adds rate headroom (25 rps) and support, not lower unit cost

Filings & research corpus

Market Intelligence over long documents, chunked caller-side

500K docs/mo × ~4 chunks × 1 analyzer2M runs
Scale $2,999 (2.5M included)$2,999/mo

→ Scale - just above the ~2M break-even, so Scale is slightly cheaper than Production + overage

Org-wide or regulated deployment

Multiple teams, data-residency or audit requirements

10M texts/mo × 1-2 analyzers10M-20M runs
Enterprise - flat annual contract, contracted per-run rateby agreement

→ Enterprise - dedicated capacity, config audits, named implementation team. Scope it with us.

Runs are billed per successfully analyzed chunk × analyzer, up to 4,096 Unicode code points per text input - "docs/mo" figures above assume the stated average chunk count per document.

INCLUDED ON PRODUCTION AND ABOVE

The Success Package

Getting thresholds wrong the first time is the single most expensive mistake in a moderation, security, or intelligence workflow. We don't let you make it alone.

30-min kickoff call

With a solutions engineer who has deployed classification systems before.

Threshold-tuning workshop

Using your own labeled data - we calibrate signals so "0.85 confidence" means what your policy needs it to mean.

Migration review

Moving from an existing classifier or LLM setup - we map where signals should differ.

30/60/90-day check-ins

Catch drift, retune thresholds, and add analyzers as your workflow evolves.

Not upsell theater - included because a team paying $799+/mo for reproducible analysis has a real workflow to protect.

Detailed plan comparison

Plan Price Included runs Overage RPS Support Success Package
Free $0/mo 5K - 2 Docs & community - Request access
Build $149/mo 50K $2.50 / 1K runs 10 Email · within 1 business day - Request access
Growth $349/mo 150K $2.00 / 1K runs 25 Onboarding + office hours - Request access
Production Most common at 250K+ $799/mo 500K $1.50 / 1K runs 50 Named customer engineer Included Request access
Scale $2,999/mo 2.5M $0.90 / 1K runs 200 CSM + TAM Included Request access
Enterprise By agreement By agreement By agreement By agreement Dedicated team Included Contact Sales →

Configuration on every plan - what each term means

Recorded
Every successful analysis result carries a verification.configuration ID, fingerprint, and signature. All plans, always - this is how the API works, not an entitlement.
Pinned
Send detection_config to freeze which configuration your requests run against - an ID is valid only for the analyzer set that issued it. Available on every plan, including Free. See Configuration pinning for lifecycle states and error behavior.
Supported
How long a pinned configuration stays replayable. Replay an analysis using its original input and supported configuration ID - availability follows the published support policy; on Enterprise it is a contractual commitment for the life of the agreement. A security-suspension event can pause execution of an affected configuration while verification of its receipts continues - a suspended configuration fails explicitly and is never silently replaced by a different detector.
See all differences - analytics tier, metadata history, API keys
Plan API keys Pin a configuration? Analytics Metadata history
Free 1 Yes Basic 90 days
Build 3 Yes Standard 90 days
Growth 8 Yes Advanced 90 days
Production 15 Yes Advanced 90 days
Scale 50 Yes Full 90 days
Enterprise By agreement Yes Full By agreement

Metadata history = operational request metadata. Submitted text is not retained on any plan.

Pricing FAQ

The top answers are also inline - hover the ? markers on the cards.

What counts as an analyzer run?

One text analyzed by one analyzer, up to 4,096 Unicode code points. A request can include at most 2 analyzers, so a single request produces at most 2 analyzer runs. Longer documents are split into chunks by the caller - each successfully analyzed chunk consumes runs.

Can I prevent overage charges?

Yes. Paid plans include automatic overage by default so production traffic is never cut off, but you can configure a monthly spend cap. Once the cap is reached, new analysis requests are rejected with HTTP 429 until the next billing month or until you raise the cap. Requests already in-flight when the cap trips finish and are billed - a cap stops new work, it doesn't kill running work. Rejected calls are not stored or queued - resubmit them yourself after the next cycle or after raising the cap.

Are failed requests billed?

No. Failed analyzer executions and valid idempotent retries are not billed. You only pay for successful analyzer runs.

Is my submitted text retained?

No. Submitted text is not retained on any plan. Plans differ only in how long operational request metadata (timestamps, configuration IDs, usage counts) is kept.

Currency, taxes, and annual billing?

All prices are US dollars, shown exclusive of any applicable taxes - taxes, where they apply, are assessed by the payment provider at checkout and shown on your invoice. Annual billing is the yearly price shown on each card: 12 months at 15% off the monthly rate, billed once. Overage is metered monthly and invoiced at your plan's published per-1,000 rate regardless of billing interval.

43 languages platform-wide - 8 to 41 per analyzer

Every plan includes every language each analyzer supports. A check means the analyzer evaluates that language today; a dash means it does not.

Language T&S trust-safety · 16 langs Cyber cybersecurity · 15 langs Discrimination discrimination · 16 langs Dating Safety dating-safety · 40 langs Fraud & Scam fraud-scam · 39 langs Crisis crisis-intelligence · 21 langs Disinformation disinformation · 16 langs Market Intelligence finance · 8 langs Insurance insurance · 40 langs Emotions emotions · 16 langs Customer Service customer-service · 18 langs Wellness wellness · 41 langs
English en ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
French fr ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Spanish es ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
German de ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Italian it ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Dutch nl ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ - ✓
Norwegian no ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Polish pl ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Portuguese pt ✓ - ✓ ✓ ✓ ✓ ✓ - ✓ ✓ - ✓
Russian ru ✓ - ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Swedish sv ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Turkish tr ✓ - ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Ukrainian uk ✓ - ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Danish da - ✓ - ✓ ✓ ✓ - - ✓ - - ✓
Finnish fi ✓ ✓ ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Latvian lv ✓ - ✓ - - ✓ ✓ - - ✓ ✓ -
Lithuanian lt ✓ - ✓ ✓ ✓ ✓ ✓ - ✓ ✓ ✓ ✓
Arabic ar - - - ✓ ✓ - - - ✓ - ✓ ✓
Hebrew he - ✓ - - - ✓ - - - - - -
Hindi hi - - - ✓ ✓ - - ✓ ✓ - ✓ ✓
Japanese ja - ✓ - ✓ ✓ ✓ - ✓ ✓ - ✓ ✓
Korean ko - ✓ - ✓ - ✓ - ✓ ✓ - - ✓
Chinese zh - ✓ - ✓ ✓ ✓ - ✓ ✓ - ✓ ✓
Egyptian Arabic arz - - - ✓ ✓ - - - ✓ - - ✓
Belarusian be - - - ✓ ✓ - - - ✓ - - ✓
Bulgarian bg - - - ✓ ✓ - - - ✓ - - ✓
Bengali bn - - - ✓ ✓ - - - ✓ - - ✓
Czech cs - - - ✓ ✓ - - - ✓ - - ✓
Greek el - - - ✓ ✓ - - - ✓ - - ✓
Hausa ha - - - - ✓ - - - - - - ✓
Croatian hr - - - ✓ ✓ - - - ✓ - - ✓
Hungarian hu - - - ✓ ✓ - - - ✓ - - ✓
Indonesian id - - - ✓ ✓ - - - ✓ - - ✓
Marathi mr - - - ✓ ✓ - - - ✓ - - ✓
Nigerian Pidgin pcm - - - ✓ ✓ - - - ✓ - - ✓
Romanian ro - - - ✓ ✓ - - - ✓ - - ✓
Slovak sk - - - ✓ ✓ - - - ✓ - - ✓
Serbian sr - - - ✓ ✓ - - - ✓ - - ✓
Swahili sw - - - ✓ ✓ - - - ✓ - - ✓
Tamil ta - - - ✓ - - - - ✓ - - ✓
Telugu te - - - ✓ ✓ - - - ✓ - - ✓
Urdu ur - - - ✓ ✓ - - - ✓ - - ✓
Vietnamese vi - - - ✓ ✓ - - - ✓ - - ✓

Coverage grows as language packs graduate - check an analyzer's docs page for its per-language evidence.

Start free. Upgrade when the workflow proves itself.

Test Sentiqra with 5K free analyzer runs. No card required.

Request access Contact Sales