MIGRATION GUIDE

Migrate from LLM moderation to Sentiqra.

Prompted LLMs are flexible but non-deterministic. When you need reproducible decisions, structured taxonomies, or audit trails - it's time to migrate.

Before and after

Before - LLM Moderation

const response = await openai.chat.completions.create({
  model: "gpt-4o",
  messages: [{
    role: "user",
    content: `Analyze this text for toxicity: "${text}"`
  }]
});

// Response: free-form text, non-deterministic
// "This text contains harassment and insults."

After - Sentiqra

const response = await sentiqra.analyze({
  text: "You're an idiot and I hope you fail.",
  language: "en",
  analyzers: ["trust-safety"]
});

// Response: structured signals with config ID
// {
//   "signals": [
//     { "code": "insults_derogation", "score": 0.86, ... },
//     { "code": "insult", "score": 0.99, ... }
//   ],
//   "verification": { "configuration": "sqcfg_ZZW77NM9", ... }
// }

Field mapping

LLM Moderation	Sentiqra	Notes
Free-form text response	signals[].code	Structured signal codes, not free-form labels
-	signals[].score	Per-indicator signal-match score (0-1)
-	signals[].level	Category vs label - hierarchical taxonomy
-	verification.configuration	Immutable config ID - enables reproducibility
-	verification.fingerprint	SHA-256 fingerprint for audit trail

Migration checklist

Map your prompt categories to Sentiqra signal codes

Sentiqra's Trust & Safety taxonomy includes structured signal codes for common moderation categories. Map your prompt's categories to these codes.

Update your API client

Replace the LLM client with Sentiqra's REST API. Use the code examples above as a starting point.

Test with your production data

Run 20 representative samples through both systems. Compare outputs side by side.

Update your policy mapping

Map Sentiqra's signal codes to your enforcement rules. Sentiqra's structured taxonomy allows more precise policy mapping than free-form LLM labels.

Enable verification receipts

Store the configuration ID and fingerprint from each successful analysis response. This enables reproducibility for appeals and audits.

Ready to migrate?

Start with 5,000 analyzer runs per month. No credit card required.

Request access Try the Playground