New · Datpaq Profanity API

Clean conversations.
No black-box moderation.

Datpaq's Profanity API combines speed, accuracy, and explainability. Filter toxic language in chat apps, games, and enterprise platforms without unpredictable AI behavior or overzealous censorship.

Start freeView API docs
POST /v1/profanity/analyze
"You are so damn rude."
Severity: 0.82 · insult
Directed at user
No negation
Token insight
“damn” · profanity · intensifier
→ severity 0.78
Explainable classification – no black-box AI.
Designed for audit logs & compliance reviews.

Profanity API

We built our Profanity API for modern real-time products: high volume, low latency, and zero tolerance for surprises. Every rule is explicit. Every decision is explainable.

Blazing Fast
Up to 100,000 lookups per second with an in-memory Redis-backed architecture designed for real-time messaging.
Remarkably Accurate
27,640 curated entries across English and Spanish with severity scores, categories, tokenization, metadata, and offensive flags.
Smart Context Awareness
We don’t just match words — we understand how they’re used, reducing noisy false positives and missed edge cases.
Built for Compliance
Explainable, auditable decisions that fit PCI environments, GDPR requirements, and regulated organizations.
Multi-Language Ready
English + Spanish out of the box with a roadmap for additional languages as your audience expands globally.
Easy to Integrate
Simple HTTP APIs that plug into your chat systems, moderation dashboards, gaming backends, EdTech platforms, and more.
Smart Context Awareness
We don't just look at words — we understand how they’re used.

Our engine looks at grammatical structure and message-level context to reduce false positives and catch subtle abuse. We detect:

  • Negation (“not a jerk” vs “you are a jerk”)
  • Quoted text and “just quoting” behavior
  • Sarcasm cues and intensifiers (“so damn rude”)
  • Self vs. directed insults
  • Softening language (“kinda dumb”, “a bit harsh”)
  • Stacked phrases that raise severity
Built for Compliance
Designed for PCI, GDPR, and regulated organizations.

Every classification includes severity, category, and rationale-friendly data so your teams can build detailed audit trails.

  • PCI-aware architecture and data handling patterns
  • Consistent outputs for legal & compliance sign-off
  • No opaque model updates that change behavior overnight

In other words: predictable, auditable, enterprise-safe.

Easy to Integrate

Use Datpaq in chat systems, moderation dashboards, parental filters, customer support tools, gaming backends, EdTech, streaming, and more.

API Endpoints
Consistent, JSON-based APIs that scale from prototypes to billions of calls.
/lookup
Identify a single word or token and retrieve severity, category, and metadata.
/bulk-lookup
Send hundreds of terms at once for efficient batch processing.
/analyze
Score full messages with context: negation, sarcasm cues, intensifiers, and more.
/severity-scale
Understand our rating system so product, trust & safety, and legal are aligned.
Example: Analyze a message
Simple REST request with clear, structured response.
Request
POST https://api.datpaq.com/v1/profanity/analyze
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "text": "You are so damn rude.",
  "locale": "en-US",
  "mode": "chat",
  "include_tokens": true
}
Response
{
  "text": "You are so damn rude.",
  "severity_score": 0.82,
  "category": "insult",
  "tokens": [
    {
      "value": "damn",
      "offset": [12, 16],
      "severity": 0.78,
      "flags": ["profanity", "intensifier"]
    }
  ],
  "context": {
    "directed_at_user": true,
    "negated": false,
    "intensifiers": ["damn"],
    "softeners": []
  }
}

Scales with You

Kubernetes-based infrastructure with autoscaling lets you start with 10 requests a day and grow to 10 million. Your traffic, your rules.

Chat & Messaging
In-app chat, support tools, and community platforms that need real-time filtering without breaking the flow.
Gaming & Esports
Protect players from harassment and toxicity while preserving fun, fast-paced banter.
EdTech & Kids Apps
Keep learning environments safe and age-appropriate with explainable filtering policies.
Customer Communities
Guard forums, reviews, and user groups to maintain a healthy, brand-safe environment.
Streaming & Creators
Moderate live chat and comments without introducing high latency or unpredictable AI behavior.
Enterprise & Regulated
Support internal communications, regulated industries, and audit-heavy workflows with repeatable classifications.

Frequently Asked Questions

Still deciding how to approach moderation? Here are answers to some common questions from product, trust & safety, and engineering teams.

Why use a dedicated Profanity API instead of manual moderation?
Manual review doesn’t scale and burns out teams. A dedicated Profanity API gives you instant, consistent decisions so moderators can focus on nuanced edge cases and appeals, not routine screening.
How is Datpaq different from AI-only moderation tools?
Many AI-only filters are opaque and hard to audit. Datpaq combines curated dictionaries, deterministic logic, and context-aware scoring so every decision is explainable, repeatable, and safe for compliance-heavy environments.
Will the filter break legitimate conversation?
Our context engine accounts for negation, quoted text, sarcasm cues, intensifiers, and softeners. That reduces false positives and keeps your chat feeling human, not over-policed.
Can I tune strictness per product or region?
Yes. Severity scores, categories, and metadata let you create different policies for kids, teens, and adults, or tailor your rules by locale and product line.

Ready to ship cleaner conversations?

Drop our Profanity API into your stack and go live in an afternoon. Start with an API key, wire it into staging, and promote to production when you're ready.

Start freeView Documentation