An API that knows which animal is in the picture
Built on millions of camera-trap images.
250 free detections every month. No card needed.
curl -X POST \ https://api.animaldetect.com/v1/detect \ -H "Authorization: Bearer $ANIMAL_DETECT_KEY" \ -F "image=@ridge-camera-0412.jpg" \ -F "country=DNK"{ "detections": [ { "label": "red deer", "confidence": 0.97, "taxonomy": { "family": "cervidae" }, "bbox": [0.50, 0.50, 0.19, 0.30] } ], "free_quota_remaining": 238}Four endpoints, one job: know what your cameras saw
Species identification
Trained on millions of camera-trap images. Tell us the country and the API narrows its species list to what actually lives there.
Empty-image filtering
Most camera-trap images are empty. The filter endpoints split animal from no-animal, so you only look at images that matter.
Batch processing
Queue thousands of images as jobs and fetch results when they are ready. Same accuracy, built for volume.
Camera-metadata OCR
Read the temperature, date, and time stamps burned into camera-trap frames as structured JSON.
JSON you can build on
Every detection comes back as data: the species with its full taxonomy, a confidence score, and a bounding box. Feed it to your own maps, alerts, or databases.
Simple pricing
250 detections free every month. After that, €1.50 per 1,000 images. Every endpoint, one meter.
First detection in minutes
- 1
Sign up
Free account, no card.
- 2
Create a key
One click in your dashboard.
- 3
Send an image
One HTTP request, or try it in the playground first.
Questions
Yes. Every organisation gets 250 free detections each month, batch jobs included. No card needed to sign up or to use the free quota.
Wild and urban animals in photos and camera-trap images. You get the species, a confidence score, and a bounding box for every animal found. Species lists are country-aware and cover 195 countries.
Create a job, upload the image or point us at a URL, and fetch the result when it is ready. Batch jobs draw from the same free monthly quota first.
Images are processed for detection and stored only as long as your job or request needs. They are never used to train models without your agreement.