A text moderation API that does more than score.
One endpoint routes every piece of text — comments, chat, messages, reviews — to OpenAI, Mistral, Azure, Llama Guard and more. Pre-filtering, custom rules, dual thresholds and a DSA-ready audit trail included.
Free self-hosted · No credit card · OpenAPI reference
Moderate any text, one request
Send text to one endpoint. Pick the provider per request or let the smart router decide. Scores come back standardized, with thresholds you control.
- ✓10 categories, 120+ languages — hate, harassment, sexual, violence, self-harm, spam, PII and custom LLM categories.
- ✓Pre-filter — block obvious spam and forbidden words before any billable provider call.
- ✓Dual thresholds — auto-approve, grey zone → human review, auto-reject.
- ✓Batch up to 10K texts — same schema as single-text calls.
# Moderate any text — comments, chat, messages, reviews
curl -X POST https://api.openmoderation.com/v1/moderation \
-H "Authorization: Bearer om_key_..." \
-H "Content-Type: application/json" \
-d '{
"input": "user comment here",
"model": "openai/omni-moderation-latest",
"categories": ["hate", "harassment", "sexual", "violence", "self-harm", "spam"],
"threshold": 0.5
}'
# Response (standardized across providers)
{
"flagged": false,
"results": [{ "category": "harassment", "score": 0.03, "flagged": false }],
"usage": { "provider": "openai", "latency_ms": 212 }
}Built for every text moderation use case
One API, every surface where users type.
Comments & reviews
Auto-hide hate speech, spam and PII leaks in comments, reviews and forum posts before they go live.
Chat & messaging
Real-time text moderation across multiple languages with sub-250ms latency and automated actions.
AI app prompts
Filter both user inputs and LLM outputs. Use LLM-based moderation for custom, evolving categories.
Marketplace listings
Flag spam listings, scams and prohibited content in titles and descriptions at ingest time.
Single-provider API vs OpenModeration
A raw toxicity score is not a moderation system.
| Capability | Single-provider text API | OpenModeration |
|---|---|---|
| Providers | 1 (lock-in) | 10 — OpenAI, Mistral, Azure, Llama Guard, Perspective, Hive… |
| Custom rules | DIY | Forbidden words, thresholds, LLM charter — no code |
| Human review | DIY | Built-in review queue with context and assignation |
| Audit trail | API logs only | Full DSA export — AI + human decisions |
| Pricing | Per call, every call billed | Free self-hosted · managed from €99/mo · pre-filter saves 10-30% |
Text moderation API — FAQ
What is a text moderation API?
A text moderation API scores user-generated text (hate, harassment, sexual, violence, spam, PII…) so your app can auto-approve, flag or reject it. OpenModeration unifies 10 providers behind one endpoint with a standardized response format, custom rules and an audit trail.
Which text moderation API should I choose?
Single-provider APIs (OpenAI, Azure, Mistral) are fast to start but lock you into one accuracy profile and one price. OpenModeration routes each request to the best provider — OpenAI for English toxicity, Mistral for EU languages, LLMs for custom categories — and adds pre-filtering, rules and a review dashboard.
How much does a text moderation API cost?
OpenAI charges ~$0.10 per 1K short texts, Mistral €0.24/1K requests, Azure $0.50-$1.00/1K records. OpenModeration is free to self-host — managed cloud starts at €99/month for 10K requests, and pre-filtering cuts 10-30% of billable calls.
Can I define custom moderation rules?
Yes. Three tools, no code: forbidden words (instant block), category thresholds (auto-approve / human review / auto-reject), and a plain-language community charter enforced by LLMs for custom categories like misinformation or grooming.
Can I migrate from openai text-moderation-latest?
Yes. Replace openai.moderations.create with one POST /v1/moderation call using model openai/omni-moderation-latest. Keep your OpenAI key (BYOK), then add fallback providers without rewriting code.
Is every text moderation decision logged?
Yes. Every AI and human decision keeps provider scores, thresholds, reviewer and timestamp. Export the full audit trail in one click for DSA transparency reports.
Start moderating text in minutes.
Get an API key, copy the curl example, and moderate your first text today — with rules, review queue and audit trails from day one.
No credit card required
Questions? contact@openmoderation.com