API

One request. A finished card.

Send a photo and a background — get back a processed photo. Your own fields, such as a listing ID, come back with it.

You don't need to figure this out yourself. Have your own developers? They can wire it all up from this page. No developers? We'll connect it for you: send us access to your site or CRM, and we'll handle the rest and show you the result on your own photos.

Flexible controlbackground, plate and car size — set per batch
10,000photos in a batch, volume on request
No redrawthe body stays as shot
curl -X POST https://auto.happyin.app/v1/frames \
  -H "Authorization: Bearer $AUTO_API_KEY" \
  -F "image=@IMG_4412.jpg" \
  -F "background=white-showroom" \
  -F 'metadata={"listing":"A-4412"}'
from happyin_auto import Client

client = Client()             # key from AUTO_API_KEY
frame = client.process_frame(
    "IMG_4412.jpg",
    background="white-showroom",
    download_to="IMG_4412_showroom.png",
)
print(frame["status"])        # done, or failed + reason
import { writeFile } from "node:fs/promises";
import { Client } from "@happyin/auto";

const client = new Client();  // key from AUTO_API_KEY
const frame = await client.processFrame("IMG_4412.jpg", {
  background: "white-showroom",
});
if (frame.status === "done") {
  const png = await client.download(frame.image_url!);
  await writeFile("IMG_4412_showroom.png", png);
}
The response usually arrives in the same request. If processing takes longer, you get a 202 and the result by polling or webhook. For batches, use /v1/batches.SDKs for Python and Node: batches, polling and webhooks in one client. For now we hand them over when we connect you; PyPI and npm are coming soon.
// what comes back
{
  "ok": true,
  "frame": {
    "id": "frm_6f1d0c9a2b8e4f7a9c3d5e1b0a2f4c6d",
    "status": "done",
    "image_url": "https://…/IMG_4412.png?X-Amz-Expires=600&…",
    "charged": true,
    "metadata": { "listing": "A-4412" }
  }
}

Five methods

POST/v1/framesOne photo → a processed photo. Usually in the same request.
POST/v1/batchesUp to 10,000 photos in one job, volume on request. A webhook fires when it's done.
GET/v1/frames/{id}A photo's status and the link to the finished file.
GET/v1/backgroundsThe background library. Your own background goes as a file with the processing request.
POST/v1/webhooksSubscribe to events: batch done, photo done.

Coming soon: your own scenes in the background library.

Getting connected

Three ways to start — with developers or without

Three ways to get started. Pick whichever fits — the first 30 frames are free on any of them.

Working through an AI agent? ChatGPT, Claude Code, Cursor and other MCP clients connect to the same API through MCP. How to connect an AI agent

01 · no developers needed

We connect it

Leave a request, we call you, take access to your site or CRM, and set up the export ourselves. You just review the result and say which background you like.

We set it up
02 · no developers needed

Just the dashboard

Nothing to connect: upload a folder of photos, pick a background, download the finished ones. Works right in the browser — anyone on your team can run it.

No integration
03 · has developers

API and a key

One POST request, three parameters, the response usually in the same connection. Python and Node SDKs. A sandbox key (free, results watermarked) arrives by email after you send a request and confirm your address.

From this page
The response

Not just a picture. Data too.

Along with the photo we return its status and, if a photo couldn't be processed, the reason; such a photo isn't charged. Your own fields from metadata, such as a listing ID or VIN, come back in the response and in the webhook.

How processing works →
{
  "ok": true,
  "frame": {
    "id": "frm_6f1d0c9a2b8e4f7a9c3d5e1b0a2f4c6d",
    "idx": 0,
    "name": "IMG_4412.jpg",
    "status": "done",
    "outcome": "accepted",
    "reason": null,
    "step": null,
    "kind": "exterior",
    "options": null,
    "mode": "full_blend",
    "shot_kind": null,
    "effective_view": null,
    "view_level": null,
    "refits_left": null,
    "brand": null,
    "attempts": 1,
    "charged": true,
    "result": true,
    "redo_in_progress": false,
    "sent": true,
    "finished_at": "2026-10-14T09:12:41.000Z",
    "vehicle": {
      "status": "unconfirmed",
      "make": { "code": "skoda", "text": "Škoda", "source": "estimate", "confidence": 0.93 },
      "model": null,
      "body": { "code": "liftback", "source": "estimate", "confidence": 0.97 },
      "dimensions_mm": { "length": 4690, "width": 1830, "height": 1470, "source": "estimate", "typical_error_pct": 4 },
      "ref": null,
      "vin": null,
      "estimate": { "model_version": "carid-2026-10", "make": [["skoda", 0.93]], "body": [["liftback", 0.97]], "mismatch": [] },
      "confirmed_at": null
    },
    "batch_id": "bat_0b7e2d4c9a1f4e6b8d3c5a7f9e1b2d40",
    "image_url": "https://…/IMG_4412.png?X-Amz-Expires=600&…",
    "image_expires_at": "2026-10-14T09:22:41.000Z",
    "metadata": { "listing": "A-4412" }
  }
}