Developer API • Coming Soon

Power Your Apps with
AI Detection API

We are currently finalizing our REST API for public release. Gain access to sentence-level analysis, probability scores, and multi-model detection with a single endpoint.

Join 2,400+ developers on the waitlist

detect_api.js

01 const detector = new GPTZero({

02 apiKey: 'gz_live_xxxxxxxx'

03 });

04 await detector.analyze(text);

Waitlist Only

Latency

< 150ms

Enterprise-Grade Performance

Built for scale, engineered for accuracy.

Model Agnostic

Detect content from ChatGPT (GPT-5.5, GPT-5.4), Claude Fable 5, Gemini 3.5, Llama 4, and more with high precision.

Sentence Level

Don't just get a score—identify exactly which parts of a document were likely AI-generated.

Global Edge

Low-latency responses served via our global edge network, perfect for high-traffic apps.

REST API Technical Specifications

Integrate high-accuracy AI classification directly into your CMS, editor, or LMS workflow with our simple JSON over HTTPS endpoints.

Core Endpoint

Submit text documents to get immediate prediction scores. The endpoint returns a general probability score for the document and identifies AI-generated patterns on a sentence-by-sentence basis.

POST /v1/detect

Required Headers

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

JSON Body Parameters

Field	Type	Description
document	string	The raw text to check. Min 150 characters, max 100,000 characters.
model_version	string	Optional. Set to "latest" or "v1.2" for stable sentence scanning.

JSON Response Payload 200 OK

{
  "status": "success",
  "data": {
    "ai_score": 0.992,
    "perplexity": 12.4,
    "burstiness": 3.8,
    "sentences": [
      {
        "id": 1,
        "text": "AI technologies provide rapid analysis of information.",
        "class": "ai"
      },
      {
        "id": 2,
        "text": "I drafted this essay late last night on my laptop.",
        "class": "human"
      }
    ]
  }
}

Rate Limits and Auth Guidelines

Secure Authentication

Generate your API keys directly inside the developer dashboard. All requests must utilize TLS 1.3 encryption and include your authorization token. Never share your key or hardcode it inside client-side JS applications.

Throttling Tiers

To ensure high system uptime, standard developer accounts are rate-limited to 60 requests per minute. Enterprise licenses feature customized quotas, high throughput parameters, and dedicated server capacity.

Developer API FAQs

Which models are currently identified by the API?

Our system detects outputs generated by ChatGPT, GPT-4, GPT-5, Gemini 3.5, Claude Fable, and Llama 4. We release update packages within days of new model releases.

Does the API require plain text or can it scan file uploads?

The baseline endpoint takes plain text strings. For file inputs (PDF, Word, TXT), we suggest reading the file contents on your server and passing the raw text to the payload.

Can we customize the threshold for AI flags?

Yes. Since our endpoint provides the exact probability score (from 0.0 to 1.0) and sentence-level labels, you can adjust your application limits to fit your specific policy.

How can we secure an early access key?

Submit your work email in the waitlist form above. We are distributing early access keys in weekly batches based on integration case volume.

Be First in Line

We're onboarding developers in batches. Join the waitlist today to secure your spot and receive early-bird pricing.

No Credit Card Req.

Instant Notification

Power Your Apps with AI Detection API