Moderate text content

Return harmful-content categories and confidence scores for supplied text. Served by text/moderation/google. Text is limited to 10,000 characters. Billed from the returned USD cost, including cached zeroes. Missing or invalid cost requires reconciliation. Interpret individual items[].likelihood_score values by category. The native nsfw_likelihood_score aggregate can include topical categories such as Health or Legal and is not a calibrated probability of dangerous content.

Input

Input for text/moderationRequired · input

Price

Usage based, at the cost the provider reports for the call

Billing

Billed at the cost the provider reports for the call, never above $0.1. Your agent sees the maximum before running.

Price checked

3 November 2026

Your agent reads the schema and the price with inspect_tool before it runs the tool.

MCP call

Copy

cURL command

Copy

TypeScript

Copy

Tool ID

Copy

• Anonymize personal details in textEden AIUsage based
• Assess whether text may be AI generatedEden AIUsage based
• Check spelling and grammarEden AIUsage based
• Extract named entities from textEden AIUsage based
• Extract topics from textEden AIUsage based
• Find potential text plagiarismEden AIUsage based
• Compare all 39