Developer API · Available now

AI Text Detector API

Q: What does the AI Text Detector API return?

It returns a document-level result and AI score, sentence-level items with their own prediction and score, execution details, word count and remaining credits.

Q: What field is required in a request?

The content string is required. callback_url, poll and webhook_secret are optional fields for asynchronous delivery.

Q: How is the binary AI result determined?

The API returns result "ai" when the document ai_score is greater than 0.5; otherwise it returns "human". Keep the numeric score and sentence-level items available for review.

Q: Where can I review current text API limits?

The detector reference documents the current request shape and response fields. Account-specific access and limits are available in the developer platform.

Q: How long are asynchronous results available?

Completed asynchronous results can be retrieved from the Jobs API for one hour.

Send text and receive a document-level AI score with sentence-level results that show which passages need a closer look.

Create an API key ↗ Read the API reference ↗

1 credit / word
Sync + async
Sentence-level items

POST /api/detector/ 200 OK

EXAMPLE REQUEST

{
  "content": "Generative systems
  can produce fluent prose..."
}

EXAMPLE RESULT

ai_score: 0.87
result: "ai"
word_count: 18

Response anatomy

Show the score and the sentence-level results.

Each response includes the document result, sentence-level scores, execution details and remaining credits. You can present the overall score without losing the passages that contributed to it.

EXAMPLE RESPONSE · 200 OK · application/json

{
  "status": "completed",
  "task_id": "...",
  "result": "ai",
  "ai_score": 0.87,
  "items": [
    {
      "text": "Generative systems can...",
      "prediction": "ai-generated",
      "ai_score": 0.87
    }
  ],
  "word_count": 18,
  "credits_remaining": 1982
}

Sentence 01

Generative systems can produce fluent academic prose.

Likely AI 87%

Sentence 02

The reviewer still needs the source context and drafting history.

Likely human 18%

Document result 87%

Example document-level result

Delivery modes

Choose synchronous or asynchronous delivery.

With no asynchronous fields, the endpoint returns the completed result with 200 OK. Add poll, callback_url or both to receive 202 Accepted and process the result asynchronously.

01 / Sync

Immediate response

No async fields

Receive the completed result in the HTTP response with status 200.

POST → analysis → result

02 / Poll

Check job status

poll: true

Receive a task ID with status 202, then retrieve the completed result from the Jobs API.

POST → task_id → GET job

03 / Webhook

Receive a callback

callback_url: "https://..."

Send the result to your endpoint and optionally verify it with your webhook secret.

POST → 202 → callback

Text detection request

Send text with one required field.

The only required field is content. Use the optional asynchronous fields when the analysis should continue after the initial request ends.

Open request parameters →

POST /api/detector/

curl -X POST \
  https://developer-portal.proofademic.ai/api/detector/ \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "Text to analyze",
    "poll": true
  }'

01 Authorize Send a scoped key

02 Submit Send the text

03 Inspect Read the score and sentences

04 Decide Send flagged passages for review

Operational details