Moderation API is a drop-in alternative to Google's Perspective API: the same toxicity, severe-toxicity, insult, threat, and identity-attack scores, returned from a single REST endpoint in 120+ languages.
Free tier · No credit card · Migrate in < 1 hour
Protecting 1000+ product teams
The five Perspective attributes most teams use map one-to-one to Moderation API categories. Beyond that, you get coverage Perspective never offered.
| Capability | Perspective API | Moderation API |
|---|---|---|
| Toxicity | ||
| Severe toxicity | ||
| Insult | ||
| Threat | ||
| Identity attack | ||
| Profanity | ||
| Sexually explicit | ||
| Harassment | Limited | |
| Hate speech (granular) | ||
| Self-harm and suicide | ||
| Violence and weapons | ||
| Spam and scams | ||
| Conversational awareness (multi-message context) | ||
| PII redaction (emails, phone numbers, addresses) | ||
| Custom models trained on your data | ||
| Image, audio and video moderation | ||
| Languages supported | ~20 | 120+ |
| Review queue + dashboard | ||
| Webhooks and rules engine | ||
| Batch endpoints | ||
| EU data residency, GDPR, DPA | US-only | |
| Commercial SLAs and support | ||
| Migration support from vendor | Not offered | Hands-on migration help |
| Pricing | Free (sunsetting) | Free tier + paid plans |
| Status as of 2026 | Ends after 2026 | Generally available |
Perspective API capabilities reflect the public service maintained by Google Jigsaw prior to its 2026 sunset announcement. Last updated .
Field mapping reference: API reference · Get hands-on migration help
Most teams migrating off Perspective end up consolidating other tools at the same time.
Pass the last few messages along with the new one. The classifier reads the thread, not just the line, so sarcasm and slow escalations stop slipping through. Perspective scores strings in isolation, which is why they often do. More on contextual analysis.
A single call returns hate speech with granular subtypes, self-harm and suicide signals, violence, spam and scams, sexual content, and PII. Perspective gives you the first two of those.
A score on its own does not solve the problem. Borderline content lands in a moderation queue for human review. The rules engine handles the obvious cases. Webhooks pipe everything back into the system you already run.
Run inference in the EU only, sign a data processing agreement, and moderate across 120+ languages. Perspective never offered any of this for regulated European workloads.
It’s a big improvement from OpenAI. Brilliant platform for moderating across all our media types - text, audio, and images!
We're using Moderation API and super happy with it. So much so that we're doubling our quota. It's become a core part of how we keep our platform safe as we scale.

The moderation API plugin has been game-changing for my company. We no longer have to spend hours a week moderating comments on our articles. Chris has been a pleasure to work with and he’s helped us every step of the way. 100% recommend!

We needed a moderation solution that could handle 50+ languages and scale with our global market. Moderation API delivered on both.
As a CPAAS platform, we deal with messages at massive scale. Moderation API handles our peak traffic without breaking a sweat, and the custom models let us fine-tune for specific context.

Skip the long evaluation cycle. Validate parity on your own historical data, then flip the switch.
Free account, no card. Generate a key from the dashboard and save it as MODERATION_API_KEY.
Replace the Perspective analyze call with a POST to /v1/moderate/text. The request body is just { value }.
Read flagged and labels[] from the response. Reuse your existing thresholds — both APIs return 0 to 1 scores.
Everything we get asked about migrating off Perspective. If something is missing, talk to us.
Google Jigsaw announced on the official Perspective site that the API will no longer be in service after 2026. The team also stopped allowing new usage and stopped approving requests for increased quotas roughly one month after the announcement. Existing integrations keep working through 2026, but every team that depends on toxicity scoring needs a migration plan.
In Jigsaw's announcement, the stated reason is that AI capabilities have evolved and there is now less demand for a standalone tool focused on this area. Jigsaw's mission is shifting toward new challenges, so the team made the decision to sunset the API rather than continue investing in it.
No. The sunset email from Jigsaw says directly that they will not provide migration support. That is part of why we built this page. If you want help porting an existing integration, talk to us and we will sit down with you.
For the attributes most teams actually use — toxicity, severe toxicity, insult, threat, identity attack, profanity, sexually explicit — yes. The request and response shapes are different, but the conceptual mapping is one-to-one. Most migrations are a single-file change.
Yes. Perspective scores a single string in isolation, which means sarcasm and slow-burn harassment often slip through. Moderation API lets you send an array of recent messages and score the most recent one with that history in mind. The classifier behaves the way a human moderator does: it reads the thread.
TOXICITY → toxic, SEVERE_TOXICITY → severe_toxic, INSULT → insult, THREAT → threat, IDENTITY_ATTACK → identity_hate, PROFANITY → profanity, SEXUALLY_EXPLICIT → sexual. Scores are floats from 0 to 1 in both APIs, so existing thresholds can be reused with minor tuning.
Yes, plus a lot more. Perspective supports around 20 languages. Moderation API supports 120+, with consistent category coverage across the long tail rather than English-only depth.
Under an hour for a single-service integration: swap the endpoint, update the auth header, and remap the response fields. Larger systems with multiple call sites usually finish in under a day. Our team can help if you get stuck — talk to us.
Yes. Sign up for a free account and you get enough quota to validate the migration end-to-end, including comparing Perspective scores against Moderation API scores on your own historical data. No credit card required.
Moderation API offers EU data residency and a signed DPA, which Perspective does not. We are GDPR compliant and run optional EU-only inference for customers under European data protection requirements.
Custom models trained on your own data, image / video / audio moderation, PII detection and redaction, a review queue and dashboard for human-in-the-loop, a rules engine, webhooks, batch endpoints, and commercial SLAs. Most teams that migrate end up consolidating other tools at the same time.