Name: SDXL Canny ControlNet LoRA API: Edge-Guided Images
Brand: magicapi
SKU: api-magicapi-sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora
Availability: InStock
Author: API Market

Shashank Agarwal

# 🖼️ Edge-Guided Image Generator API (fermatresearch SDXL Canny ControlNet LoRA)

## Overview
This model runs SDXL with a Canny edge ControlNet and LoRA support. Its developer describes it as "SDXL Canny controlnet with LoRA support" that now also supports img2img.

Send a reference `image` and a `prompt`. The model extracts the edges of the image and generates a new picture that follows them, styled by the prompt. `condition_scale` sets how strictly the edges are followed. Turn on `img2img` to also start from the reference image, with `strength` setting how much of it is replaced. Each request returns 1 to 4 images (`num_outputs`).

## Features
- **Edge-guided generation**: a Canny ControlNet keeps the outlines and composition of the reference image.
- **img2img mode**: `img2img` uses the image as both the control image and the starting image; `strength` (default `0.8`) sets how much changes.
- **Refiner**: `refine` `base_image_refiner` (default) with `refine_steps`, or `no_refiner`.
- **LoRA support**: load LoRA weights with `lora_weights` and set their influence with `lora_scale`.
- **Sampling controls**: 7 schedulers, 1 to 500 steps, `guidance_scale` 1 to 50, `negative_prompt` and `seed`.
- **1 to 4 images per request** with `num_outputs`.

## Use Cases
1. **Restyling photos**: keep the pose and layout of a photo while changing the setting or style with the prompt.
2. **Sketch to image**: turn line drawings or outlines into finished images.
3. **Consistent compositions**: generate several variations that share the same structure.
4. **Product and architecture mockups**: keep the shapes of an object or building and change materials and lighting.

## Visual Examples
### Example Input
The reference image below, with the prompt "a woman in alaska, cinematic photo", `condition_scale` `0.5`, 40 steps and the base image refiner with 20 steps. `img2img` was not set.

![Example input: the reference image](https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-input-api-market.jpg)

```json
{
  "image": "https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-input-api-market.jpg",
  "prompt": "a woman in alaska, cinematic photo",
  "refine": "base_image_refiner",
  "scheduler": "K_EULER",
  "lora_scale": 0.95,
  "num_outputs": 1,
  "refine_steps": 20,
  "guidance_scale": 7.5,
  "apply_watermark": true,
  "condition_scale": 0.5,
  "negative_prompt": "",
  "num_inference_steps": 40
}
```

### Example Output
![SDXL Canny ControlNet output: a cinematic photo of a woman in Alaska following the reference image's outlines](https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-output-4v7j2yht-api-market.jpg)

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | `An astronaut riding a rainbow unicorn` |  | Text prompt |
| `image` | string (URL) | No | none |  | Image URL (jpeg, png, gif or webp). Its Canny edges guide the generation; in img2img mode it is also the starting image |
| `img2img` | boolean | No | none |  | Use the img2img pipeline: the image is used both as the control image and as the base image |
| `condition_scale` | number | No | `1.1` | `0` to `2` | How strongly the ControlNet edges steer the result. Higher values follow the edges more closely |
| `strength` | number | No | `0.8` | `0` to `1` | Denoising strength when img2img is on. 1 discards the input image completely |
| `negative_prompt` | string | No | `""` |  | Negative prompt: what to avoid in the image |
| `num_inference_steps` | integer | No | `30` | `1` to `500` | Number of denoising steps |
| `num_outputs` | integer | No | `1` | `1` to `4` | Number of images to return |
| `scheduler` | string | No | `K_EULER` | `DDIM`, `DPMSolverMultistep`, `HeunDiscrete`, `KarrasDPM`, `K_EULER_ANCESTRAL`, `K_EULER`, `PNDM` | Sampling scheduler |
| `guidance_scale` | number | No | `7.5` | `1` to `50` | Scale for classifier-free guidance |
| `seed` | integer | No | none |  | Random seed. Leave blank for a random seed |
| `refine` | string | No | `base_image_refiner` | `no_refiner`, `base_image_refiner` | Refinement: no_refiner or base_image_refiner |
| `refine_steps` | integer | No | `10` |  | Number of refinement steps when refine is base_image_refiner |
| `apply_watermark` | boolean | No | `true` |  | Apply an invisible watermark so generated images can be identified later. Set to false to turn it off |
| `lora_scale` | number | No | `0.95` | `0` to `1` | LoRA additive scale. Applies only when LoRA weights are loaded |
| `lora_weights` | string | No | none |  | URL of LoRA weights to load. Leave blank to use the default weights |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.0039 per API unit.
- **1 API unit = 1 request with up to 3 seconds of processing time.** A request that takes longer is billed one unit per started 3-second block (for example, 7 seconds of processing is 3 units, $0.0117).
- More steps, refinement steps and images (`num_outputs`) take longer, so requests often use several units.
- **Example cost**: the example (1 image, 40 steps plus 20 refinement steps) took 9.12 seconds of processing, 4 started 3-second blocks: 4 API units = **$0.0156**.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

A polling interval of 1 second and an overall timeout of 600 seconds (10 minutes) work well.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "3bb13fe1c33c35987b33792b01b71ed6529d03f165d1c2416375859f09ca9fef",
    "input": {
      "image": "https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-input-api-market.jpg",
      "prompt": "a woman in alaska, cinematic photo",
      "refine": "base_image_refiner",
      "scheduler": "K_EULER",
      "lora_scale": 0.95,
      "num_outputs": 1,
      "refine_steps": 20,
      "guidance_scale": 7.5,
      "apply_watermark": true,
      "condition_scale": 0.5,
      "negative_prompt": "",
      "num_inference_steps": 40
    }
  }'
```

Response (201 Created):
```json
{
  "id": "4v7j2yhtn9rmt0d1cd99vdghew",
  "version": "3bb13fe1c33c35987b33792b01b71ed6529d03f165d1c2416375859f09ca9fef",
  "status": "starting",
  "created_at": "2026-10-22T11:14:56.035Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "4v7j2yhtn9rmt0d1cd99vdghew",
  "status": "succeeded",
  "output": [
    "https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-output-4v7j2yht-api-market.jpg"
  ],
  "error": null,
  "created_at": "2026-10-22T11:14:56.035Z",
  "started_at": "2026-10-22T11:17:12.350Z",
  "completed_at": "2026-10-22T11:17:21.472Z"
}
```

`output` is a list of image URLs, one per requested image (`num_outputs`).

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a `condition_scale` above 2, a `num_outputs` above 4 or a `scheduler` that is not in the list.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying and honour any `Retry-After` header. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "3bb13fe1c33c35987b33792b01b71ed6529d03f165d1c2416375859f09ca9fef", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 600000;
  const pollInterval = 1000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output;
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 10 minutes");
}

createAndWaitForPrediction({
    "image": "https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-input-api-market.jpg",
    "prompt": "a woman in alaska, cinematic photo",
    "refine": "base_image_refiner",
    "scheduler": "K_EULER",
    "lora_scale": 0.95,
    "num_outputs": 1,
    "refine_steps": 20,
    "guidance_scale": 7.5,
    "apply_watermark": true,
    "condition_scale": 0.5,
    "negative_prompt": "",
    "num_inference_steps": 40
  }).then((output) => console.log(output));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/sdxl-canny-controlnet-image-generator-fermatresearch-sdxl-controlnet-lora"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "3bb13fe1c33c35987b33792b01b71ed6529d03f165d1c2416375859f09ca9fef", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 600
    poll_interval = 1
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return prediction["output"]

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 10 minutes")


print(create_and_wait_for_prediction({
    "image": "https://blog.api.market/wp-content/uploads/2026/10/sdxl-controlnet-lora-input-api-market.jpg",
    "prompt": "a woman in alaska, cinematic photo",
    "refine": "base_image_refiner",
    "scheduler": "K_EULER",
    "lora_scale": 0.95,
    "num_outputs": 1,
    "refine_steps": 20,
    "guidance_scale": 7.5,
    "apply_watermark": True,
    "condition_scale": 0.5,
    "negative_prompt": "",
    "num_inference_steps": 40
}))
```

## FAQ
- **What is the difference between the default mode and `img2img`?**
  - By default only the edges of `image` guide the new picture. With `img2img` set to `true`, the image is also the starting point and `strength` sets how much of it is replaced (1 replaces it completely).
- **How do I keep closer to the reference outlines?**
  - Raise `condition_scale` (0 to 2, default 1.1). Lower values give the prompt more freedom.
- **Can I use my own LoRA?**
  - `lora_weights` takes a URL of LoRA weights to load, and `lora_scale` (0 to 1) sets their influence. Leave `lora_weights` blank to use the default weights.
- **How many images can one request return?**
  - 1 to 4, set with `num_outputs`. More images take longer and so use more units.
- **How is a request priced?**
  - $0.0039 per started 3 seconds of processing. The example took 9.12 seconds, so it cost 4 units ($0.0156).
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, adjust the input and try again.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


Generate images with SDXL guided by the Canny edges of a reference image, with img2img, a base image refiner and LoRA weights. Pay as you go, no monthly fee.