Image Recognition API for Rights Clearance and Brand Safety
A copyright-clean photo can still get you in trouble. A recognizable face needs a model release, a brand logo raises trademark questions, and some landmarks need property releases. These endpoints find all three, plus labels and content-safety ratings.
curl -X POST https://app.picdefense.io/api/v2/logo \
-H "X-API-TOKEN: USERID:APIKEY" \
-H "Content-Type: application/json" \
-d '{"url": "https://example.com/photo.jpg"}'{
"status": 1,
"message": "succeeded",
"error": "none",
"data": {
"logo": true,
"descriptions": "Nike, Adidas"
}
}Pay Per Call. No Subscription.
Buy prepaid credits once and spend them on any PicDefense endpoint.
| Endpoint | Credits / call | Per 1,000 calls |
|---|---|---|
POST /logoDetects brand logos and names them. | 0.03 | $2.10 – $6.00 |
POST /faceDetects whether a human face is present. | 0.03 | $2.10 – $6.00 |
POST /landmarkDetects a recognizable landmark and names it. | 0.03 | $2.10 – $6.00 |
POST /labelsDescriptive labels with confidence scores. | 0.03 | $2.10 – $6.00 |
POST /safesearchAdult, violence, racy, medical, and spoof likelihoods. | 0.03 | $2.10 – $6.00 |
Ranges run from the 100,000-credit pack down to the 100-credit pack. Credits are charged only on successful calls and never expire. See all credit packs on the pricing page.
What It Detects
Brand Logos
Names the brands visible in an image so you can review trademark use before publishing.
Faces
Flags images with people so your team can confirm a model release is on file.
Landmarks
Identifies famous buildings and places, some of which carry property-release or usage restrictions.
Labels
Describes what is in the image for tagging, search, and auto-generated alt text.
Content Safety
Rates adult, racy, violent, medical, and spoof content on a five-step likelihood scale.
Built for These Workflows
Where teams wire this endpoint into their pipelines.
Release Checks for Marketing
Before an image runs in an ad, flag faces and logos and require the matching release or approval.
Print-on-Demand Trademark Screening
Hold designs that show sports, fashion, or entertainment brands for review before they print.
Brand-Safe UGC
Combine safesearch with logo detection to keep unsafe or off-brand uploads out of public feeds.
Auto-Tagging a Media Library
Label every asset on import so editors can find images by what they show.
What You Get Back
Each endpoint is a separate call with its own response. Every response includes status, message, error, and a data object. These are the fields inside data.
POST /logo
Detects brand logos and names them.
| Field | Type | Description |
|---|---|---|
| logo | boolean | Whether any logos were found. |
| descriptions | string | Comma-separated list of detected brand names, such as Nike, Adidas. |
POST /face
Detects whether a human face is present.
| Field | Type | Description |
|---|---|---|
| face | boolean | Whether the image contains a human face. |
POST /landmark
Detects a recognizable landmark and names it.
| Field | Type | Description |
|---|---|---|
| landmark | boolean | Whether a landmark was found. |
| description | string | The landmark's name, such as Colosseum. |
POST /labels
Descriptive labels with confidence scores.
| Field | Type | Description |
|---|---|---|
| labels.labelsOnly | string[] | Detected labels, such as mountain or sky. |
| labels.labelsWithScore | string[] | The same labels with confidence, such as nature:0.98. |
POST /safesearch
Adult, violence, racy, medical, and spoof likelihoods.
| Field | Type | Description |
|---|---|---|
| safesearch.adult | string | Likelihood of adult content: VERY_UNLIKELY, UNLIKELY, POSSIBLE, LIKELY, or VERY_LIKELY. |
| safesearch.racy | string | Likelihood of racy content, on the same scale. |
| safesearch.violence | string | Likelihood of violent content, on the same scale. |
| safesearch.medical | string | Likelihood of medical content, on the same scale. |
| safesearch.spoof | string | Likelihood of spoofed content, on the same scale. |
Why Teams Choose the PicDefense Image Recognition API
Organized Around Releases
Faces, logos, and landmarks map to model, trademark, and property releases, which is how legal teams actually review images.
One Key With the Copyright Endpoints
Use the same token and credit balance as watermark detection, reverse search, and the copyright risk check.
Call Only What You Need
Each detector is its own endpoint, so you pay for the signals you use rather than a bundle.
Plain JSON Answers
Booleans and names instead of bounding-box dumps, ready for an if-statement or a review queue.
Prefer a Dashboard?
In the PicDefense platform, logo, face, and landmark detection run on every crawled image and feed into its risk score.
See how detection feeds risk scoringUse It From an AI Agent
Every endpoint is also exposed as a tool on the hosted PicDefense MCP server. Connect Claude, Cursor, or any MCP client with your existing API token.
https://mcp.picdefense.io/mcpMCP setup guidePairs Well With
One API key and one credit balance cover every endpoint.
Copyright Risk Check API
One call returns a high/medium/low copyright risk verdict plus the evidence behind it.
Learn moreWatermark Detection API
Detect visible stock and photographer watermarks and learn which agency they belong to.
Learn moreCopyright Metadata API
Pull EXIF copyright notices and read credit lines baked into the pixels with OCR.
Learn moreFrequently Asked Questions
Does the face endpoint identify who the person is?
No. It only reports whether a human face is present. It does not perform facial recognition or identify individuals.
Can I call all five detectors in one request?
Each detector is a separate endpoint at 0.03 credits. If you want face, logo, and landmark data together with a copyright verdict, the Copyright Risk Check API returns them in one call.
Does a detected logo mean the image infringes a trademark?
Not necessarily. Many uses of logos are lawful. Detection tells you where to look so a person can make the call.
How much do these endpoints cost?
Each uses 0.03 credits per successful call, roughly $2 to $6 per 1,000 images depending on your credit pack. Credits never expire.
Start Calling the Image Recognition API Today
Create an account, copy your API token from Settings, and make your first request in minutes.
PicDefense API results are automated, probabilistic signals intended to support your own review. They do not constitute legal advice and do not guarantee that an image is free of third-party rights. Consult a qualified attorney for decisions about copyright, trademark, or publicity rights.