# EditorAI

> EditorAI (https://editorai.pro) edits photos from plain-English prompts. Upload a photo, describe the change ("remove the background people", "turn this into a twilight shot", "make it look like a magazine ad"), get a finished image back in 10-40 seconds. Built for restaurants, real estate, e-commerce, salons, rentals, fitness and anyone who needs professional photos without an editor.

## For AI agents: MCP server (recommended)

Connect via MCP at:

    https://editorai.pro/mcp/<API_KEY>

The user gets this URL from the "Copy agent link" button at the top of https://editorai.pro/dashboard (or Settings → "AI Agents & API") and pastes it into Claude (Settings → Connectors → Add custom connector), ChatGPT (Settings → Connectors), or any MCP client. Transport: Streamable HTTP (POST JSON-RPC). No OAuth — the key in the URL is the auth.

Tools exposed: edit_image, bulk_edit_images, generate_image, upload_image, list_images, get_account.

## REST API

Base URL: https://editorai.pro/api/v1
Auth: 'Authorization: Bearer <API_KEY>' header. Keys start with "eai_".
All requests/responses are JSON. Endpoints are synchronous — the response contains the finished image URL.

### POST /api/v1/edits — edit an image (1 credit)
Request: { "prompt": "Remove the scooter", "image_url": "https://..." }
Multiple reference images combined into one output: { "prompt": "...", "image_urls": ["https://...", "https://..."] }
Optional: "aspect_ratio": "16:9" (defaults to matching the input image).
Response: { "id": "...", "url": "https://...", "prompt": "...", "model": "...", "status": "completed" }

### POST /api/v1/edits/bulk — same prompt on up to 10 images (1 credit each)
Request: { "prompt": "Brighten and clean up", "image_urls": ["https://...", ...] }
Response: { "total": 3, "succeeded": 3, "results": [{ "image_url": "...", "url": "...", "status": "completed" }, ...] }

### POST /api/v1/generations — create an image from text (60 credits)
Request: { "prompt": "A cozy cafe interior at golden hour", "aspect_ratio": "16:9" }

### POST /api/v1/files — add an image to the user's library (free)
By URL: { "url": "https://...", "name": "optional-name.jpg" }
By raw bytes: { "data": "<base64 or data: URI>", "name": "optional-name.jpg" }
Response: { "id": "...", "url": "https://...", "name": "...", "width": ..., "height": ... }

### GET /api/v1/files?limit=20&type=image — list the user's library (free)
Each item includes an "id" used by the endpoints below.

### GET /api/v1/files/:id — details of one file (free)
### GET /api/v1/files/:id/download — stream the raw image bytes (free; also accepts ?key=)
### DELETE /api/v1/files/:id — permanently delete a file from library + storage (free)

### GET /api/v1/me — account, remaining credits, and costs (free)

### GET fallbacks for restricted sandboxes
If your environment cannot send POST requests or blocks this domain from code execution, the core actions also work as plain GET URLs with the key as a query param (URL-encode values):
GET /api/v1/edit?key=<API_KEY>&prompt=...&image_url=... (repeat image_url to combine)
GET /api/v1/generate?key=<API_KEY>&prompt=...&aspect_ratio=16%3A9
GET /api/v1/files?key=<API_KEY> and GET /api/v1/me?key=<API_KEY>
Same JSON responses. Edit/generate take 10-40s; if your fetch times out, the job usually still completed — check the files list for the newest entry.

## Errors

Errors return { "error": { "code": "...", "message": "..." } } with HTTP 400/401/402/500.
Notable codes: "insufficient_credits" (HTTP 402 — the user needs to top up at https://editorai.pro/pricing), "image_unreachable" / "not_an_image" (the input URL must be a publicly accessible direct image link).

## Notes for agents

- Input images must be on publicly accessible URLs. If the user's photo is not, import it first via POST /api/v1/files (or the upload_image MCP tool).
- Result URLs are permanent public links — safe to show or send to the user.
- Be specific in prompts about what should change AND what should stay the same.
- Edits cost 1 credit, generation 60 credits. Check GET /api/v1/me before large batches.

## More

- Human docs: https://editorai.pro/developers
- Get an API key: https://editorai.pro/dashboard → Settings → AI Agents & API
- Pricing: https://editorai.pro/pricing