API reference · v1

SeeFood REST API

One endpoint that answers one question. Every API scan is recorded in your workspace's scan history with channel api.

Authentication

Create a key under API keys in the app and send it as a bearer token. Keys start with sf_live_, are shown once, and are stored only as a hash. Revoke a key at any time.

Authorization: Bearer sf_live_…

Classify an image

POST /api/v1/classify

Send multipart/form-data with an image field: JPEG, PNG or WebP, up to 5 MB.

curl https://seefoodsystems.com/api/v1/classify \
  -H "Authorization: Bearer $SEEFOOD_API_KEY" \
  -F image=@lunch.jpg

// Node.js 18+
const form = new FormData();
form.append("image", new Blob([await fs.readFile("lunch.jpg")]), "lunch.jpg");

const res = await fetch("https://seefoodsystems.com/api/v1/classify", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.SEEFOOD_API_KEY}` },
  body: form,
});
const scan = await res.json();
console.log(scan.is_hot_dog ? "Hot Dog" : "Not Hot Dog");

# Python
import os, requests

res = requests.post(
    "https://seefoodsystems.com/api/v1/classify",
    headers={"Authorization": f"Bearer {os.environ['SEEFOOD_API_KEY']}"},
    files={"image": open("lunch.jpg", "rb")},
)
print(res.json()["verdict"])

Response

{
  "id": "69413952-892a-4761-b412-66933a69a0db",
  "verdict": "not_hot_dog",
  "is_hot_dog": false,
  "confidence": 0.99,
  "label": "golden retriever dog sitting on grass",
  "reasoning": "The image shows a dog, not a sausage in a bun.",
  "model": "claude-haiku-5-5",
  "latency_ms": 1730,
  "hot_dog_score": null,
  "improvements": [],
  "created_at": "2026-10-17T14:04:14.360Z"
}

id	string (uuid)	Scan ID, also shown in your workspace audit log
verdict	"hot_dog" \| "not_hot_dog"	The definitive answer
is_hot_dog	boolean	Convenience flag, same as verdict === "hot_dog"
confidence	number 0–1	Model certainty in the verdict
label	string	Short description of what the image shows
reasoning	string	One-sentence explanation
model	string	Model that produced the verdict
latency_ms	integer	Inference time
hot_dog_score	integer 0–100 \| null	Quality rating, hot dogs only
improvements	string[]	What would raise the score (empty for non hot dogs)
created_at	string (ISO 8601)	When the scan was recorded

Errors

Errors return JSON shaped as { "error": { "code", "message" } }.

Status	Code	Meaning
400	missing_image / invalid_body	No `image` field, or the body isn't multipart/form-data
401	invalid_api_key	Missing, malformed or revoked API key
413	image_too_large	Image is larger than 5 MB
415	unsupported_type	File isn't a JPEG, PNG or WebP image
429	rate_limited	More than 20 scans per minute for this workspace
502	classifier_unavailable	The vision model failed; safe to retry

Limits

20 scans per minute per workspace, shared between the web app and the API.
Images up to 5 MB. Larger photos should be downscaled client-side; 1280 px on the long side is plenty.
10 active API keys per workspace.