Developer

API Reference

FractalQ lets you build visual AI pipelines and expose them as REST APIs. Connect nodes on the canvas, publish a session, and call it from anywhere — no infrastructure needed.

Base URLhttps://fractalq.com

ProtocolHTTPS only

Formatapplication/json

Quick Start

Open the dashboard and create a free account.

Build

Drag AI nodes onto the canvas and wire them together.

Publish

Hit Publish — your pipeline becomes a live REST API instantly.

Call

POST to the execute endpoint with your inputs and get outputs back.

Execute Endpoint

POST/api/node/api/{apiKey}/execute

Runs the published pipeline synchronously and returns all outputs. For pipelines that take more than a few seconds, use async: true instead.

Request

cURL

curl -X POST https://fractalq.com/api/node/api/{apiKey}/execute \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": {
      "image_url": "https://example.com/image.jpg",
      "command": "enhance and upscale"
    }
  }'

Response · 200

json

{
  "success": true,
  "outputs": {
    "result_url": "https://storage.fractalq.com/out/abc123.jpg",
    "metadata": { "scale": 4, "width": 4096, "height": 4096 }
  },
  "executionMs": 3420
}

Body Parameters

inputs*

object

Key/value pairs matching your session's input nodes.

async

boolean

Set true to run asynchronously. Returns jobId immediately.

callbackUrl

string

HTTPS endpoint to receive results when async is true.

_jobId

string

Optional idempotency key for deduplication.

Get Session Docs

GET/api/node/api/{apiKey}/docs

Returns the full input/output schema, usage examples, and environment variable for a specific session. Useful for introspection and generating client SDKs.

shell

GET https://fractalq.com/api/node/api/{apiKey}/docs

Response · 200

json

{
  "success": true,
  "apiKey": "abc123...",
  "apiUrl": "https://fractalq.com/api/node/api/abc123.../execute",
  "inputSchema": {
    "image_url": { "type": "string", "required": true, "description": "Image URL to process" }
  },
  "outputSchema": {
    "result_url": { "type": "string", "description": "Processed image URL" }
  },
  "inputExamples": {
    "image_url": "https://example.com/image.jpg"
  }
}

Async Processing

Long-running pipelines (e.g. 4× upscaling) can exceed HTTP timeout limits. Pass async: true to queue the job and receive results via a callback instead of waiting for the response.

Your Server

POST /execute

FractalQ

NeuronQ AI

POST callbackUrl

cURL

curl -X POST https://fractalq.com/api/node/api/{apiKey}/execute \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": { "image_url": "https://example.com/image.jpg" },
    "async": true,
    "callbackUrl": "https://yourapp.com/api/fractalq/callback",
    "_jobId": "job-abc-123"
  }'

Your callbackUrl must be a publicly reachable HTTPS endpoint. FractalQ will POST the same response shape as a synchronous call, plus a _jobId field.

Response Codes

200

Pipeline completed. Outputs returned in response body.

202

Accepted

Job queued (async mode). Results delivered via callback URL.

400

Bad Request

Missing required inputs or malformed request body.

404

Not Found

API key not found or session not published.

422

Unprocessable Entity

Input validation failed (e.g. unreachable image URL).

429

Too Many Requests

Rate limit exceeded. Slow down requests.

500

Internal Server Error

AI processing failed. Check logs or retry.

Error Response Shape

json

{
  "success": false,
  "error": "Missing required input: image_url",
  "code": "MISSING_INPUT"
}

Node Types

Each node type performs a specific AI operation. Combine them on the canvas to build multi-step pipelines.

Image Processing

↑

IMAGE__UPSCALE

4× upscale via Real-ESRGAN. Input: image URL. Output: upscaled image URL.

≡

IMAGE__OCR

Text extraction supporting Arabic & English via PaddleOCR.

⊘

IMAGE__NSFW_GATE

NSFW detection & filter. Blocks pipeline if explicit content is detected.

◈

IMAGE__WATERMARK

Detects and localizes watermarks. Returns bounding boxes and confidence.

I/O

→

INPUT__IMAGE

Accepts an image_url string. Entry point for image-based pipelines.

←

OUTPUT__

Pipeline output sink. Collects and returns all upstream results.

Authentication

The execute endpoint is public — no Authorization header required. Access is controlled entirely by the API key embedded in the URL. Treat your API key like a password.

Keep it secret

Never expose your API key in client-side code or public repos.

Rotate anytime

Regenerate your key from the dashboard. Old keys are revoked immediately.

Scope by pipeline

Each published session gets its own key — scope access per use case.

shell

# Your API key is part of the URL — no headers needed
POST https://fractalq.com/api/node/api/YOUR_API_KEY/execute

Questions? dev@internalt.com

Terms Privacy Contact