Moderations

Classify text for potentially harmful content using the OpenAI-compatible moderation API with BYOK support.

Classify text for potentially harmful content using the content moderation API. Requests are proxied to OpenAI's moderation endpoint with automatic BYOK key resolution.

POST /v1/moderations
Auth: Authorization: Bearer <token>
Drop-in compatible with the OpenAI Moderations API.

Quick start

curl -sS https://api.xantly.com/v1/moderations \
  -H "Authorization: Bearer $XANTLY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "I love programming and building useful tools."
  }'

Request body

Field	Type	Required	Description
`input`	`string \| array<string>`	Yes	Text to classify. Pass a string for single input, or an array for batch classification.
`model`	`string`	No	Moderation model to use (e.g. `"omni-moderation-latest"`). Defaults to the latest available model.

Response body

The response follows the OpenAI moderation response format:

{
  "id": "modr-abc123",
  "model": "omni-moderation-latest",
  "results": [
    {
      "flagged": false,
      "categories": {
        "sexual": false,
        "hate": false,
        "harassment": false,
        "self-harm": false,
        "sexual/minors": false,
        "hate/threatening": false,
        "violence/graphic": false,
        "self-harm/intent": false,
        "self-harm/instructions": false,
        "harassment/threatening": false,
        "violence": false
      },
      "category_scores": {
        "sexual": 0.000012,
        "hate": 0.000003,
        "harassment": 0.000008,
        "self-harm": 0.000001,
        "sexual/minors": 0.000001,
        "hate/threatening": 0.000001,
        "violence/graphic": 0.000002,
        "self-harm/intent": 0.000001,
        "self-harm/instructions": 0.000001,
        "harassment/threatening": 0.000002,
        "violence": 0.000004
      }
    }
  ]
}

Field	Type	Description
`id`	`string`	Moderation request identifier.
`model`	`string`	Model used for classification.
`results`	`array`	One result per input string.
`results[].flagged`	`boolean`	`true` if any category is flagged.
`results[].categories`	`object`	Boolean flags per content category.
`results[].category_scores`	`object`	Confidence scores per category (`0.0`-`1.0`).

Code examples

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["XANTLY_API_KEY"],
    base_url="https://api.xantly.com/v1",
)

response = client.moderations.create(
    input="This is a test message for content moderation.",
)

result = response.results[0]
print(f"Flagged: {result.flagged}")
for category, flagged in result.categories.items():
    if flagged:
        print(f"  {category}: {result.category_scores[category]:.4f}")

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.XANTLY_API_KEY,
  baseURL: "https://api.xantly.com/v1",
});

const response = await client.moderations.create({
  input: "This is a test message for content moderation.",
});

const result = response.results[0];
console.log(`Flagged: ${result.flagged}`);

BYOK support

The moderation endpoint automatically resolves your organization's BYOK OpenAI key. If no BYOK key is configured, the platform key is used as a fallback.

Errors

HTTP	`error.type`	Typical trigger
`400`	`invalid_request_error`	Missing `input` field.
`401`	`authentication_error`	Missing or invalid Bearer token.
`500`	`provider_error`	No OpenAI API key configured, or upstream error.

Next steps

Chat Completions, Main inference endpoint
Models, List available models