Image face crop

POST /api/image-face-crop3 quota units per call · cache hits free

Detect a face in a photo and return a tight rectangular crop centred on the face. Useful for avatar generation from full-body photos.

Detects a face in the input photo and returns a tight crop centred on it. Configurable padding around the face. Uses a deep-learning face detector (more reliable than classic Haar cascades, especially for non-frontal faces and partially-occluded faces). Useful for generating avatar / profile-picture crops from full-body or group photos.

Inputs

Name	Type	Default	Description
image*	file	—	Source image bytes.
padding	number (0…2)	0.4	Padding around the face bounding box, as a fraction of face size. 0 = tight; 0.4 = comfortable avatar.
aspect	string	"square"	Output aspect ratio. "square" (default), "original" (preserve face box), or a numeric ratio like "4:5".
outWidth	number (0…4096)	0	Resize the crop to this width (height derived from aspect). 0 = keep crop's native size.
format	enum (jpeg \| png)	"jpeg"	Output encoding.
quality	number (1…100)	90	JPEG quality (ignored for PNG).
scoreThreshold	number (0…1)	0.6	confidence threshold for accepting a face.

Response

Modes: binary, stream, json. Cache: yes (24h TTL).

Large outputs stream: responses over the 4 MiB cache limit are piped through as they are produced instead of buffered. They carry x-cache-skip: size and are never cached — identical follow-up calls recompute.

Every response carries x-cache (HIT / MISS / BYPASS), x-cache-signature (stable across identical inputs), and the rate-limit headers listed below.

Quota & limits

Cost per call	3 units against the 10,000-unit monthly quota. Cache hits are free — only cache misses decrement.
Burst limit	200 requests per rolling minute, shared across all tools.
Max request size	`25 MiB` (base64 inflates file payloads ~33% — the limit applies to the decoded bytes).
Max response size	`25 MiB`
Timeout	30s wall clock.
Rate-limit headers	`X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Burst-Limit` on every response; `X-Quota-Warning` once 80% of the monthly quota is spent; `Retry-After` on 429s.

Errors

Errors are JSON with an error string; schema failures add a details object (Zod's flattened field errors).

Status	When	Body
400	Input fails the tool's schema (wrong type, missing required field, malformed base64) or the file bytes can't be parsed as the expected format.	{ "error": "invalid inputs", "details": { "fieldErrors": { … } } }
401	Missing or invalid X-Api-Key / X-Account-Id headers (or no session when called from the browser).	{ "error": "unauthorized" }
404	Unknown tool id, or calling /api/* on the website hostname instead of api.xtract.bot.	{ "error": "not found" }
413	Request body over the tool's max request size, or an image header declaring more megapixels than the tool's pixel budget.	{ "error": "…exceeds maxRequestBytes…" }
415	Content-Type isn't application/json on the json-body transport.	{ "error": "unsupported content-type; expected application/json" }
429	Monthly quota or the 200/min burst limit exhausted. Check Retry-After and the X-RateLimit-* headers.	{ "error": "monthly quota exceeded", "monthRemaining": 0, … }
5xx	Conversion engine failure. Safe to retry; if it persists, the input is hitting a bug — please report it.	{ "error": "…" }

Code samples

Built from the default example.

# Download or substitute the example input:
#   curl -O https://xtract.bot/examples/face-fixtures/face-small.png
IMAGE=$(base64 -w0 < face-small.png)

curl -X POST https://api.xtract.bot/api/image-face-crop \
  -H "Content-Type: application/json" \
  -H "Accept: application/octet-stream" \
  -H "X-Account-Id: $XTRACT_ACCOUNT_ID" \
  -H "X-Api-Key: $XTRACT_API_KEY" \
  -d '{
  "image": "'"$IMAGE"'"
}'