Detection infrastructure, not a toy
Bring multilingual, privacy-first AI detection into your product, CMS or editorial pipeline with a single endpoint.
A simple request
POST text and receive a structured verdict: an AI-likelihood score, a confidence level, the detected language, the likely generator, and sentence-level segments.
- Automatic language detection across six languages
- Model-agnostic — covers paraphrased and humanized text
- Batch endpoint for high-volume screening
- Processed in-house on GCP, no third-party sharing
curl https://api.detector.dng.ai/v1/scan \
-H "Authorization: Bearer $DNG_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"text": "Le texte à analyser…",
"language": "auto"
}'
# → {
# "ai_probability": 0.94,
# "confidence": "high",
# "language": "fr",
# "likely_model": "gpt-4o",
# "segments": [ ... ]
# } API questions
Is the API generally available?
Access is rolling out to publishers, platforms and agencies. Join the access list and we’ll reach out with credentials and limits.
Which languages does the API support?
The same six as the web app — English, Spanish, French, Italian, Portuguese and Romanian — with automatic language detection.
Where is my data processed?
In-house on Google Cloud. Submitted text is not shared with third parties and is not used to train models.
How is access scoped?
Contact us with your use case and expected volume. We will confirm availability, limits and commercial terms.
Build detection into your product
Request access and we’ll get you a key. Private by design, multilingual by default.
Contact us