Name: Magic Image Refiner API: Add Detail to Images
Brand: magicapi
SKU: api-magicapi-magic-image-refinement-upscaling-fermatresearch-magic-image-refiner
Availability: InStock
Author: API Market

Shashank Agarwal

# ✨ Image Refiner API (fermatresearch Magic Image Refiner)

## Overview
Magic Image Refiner, published by fermatresearch, adds detail to an existing image. Its developer describes it as a better alternative to SDXL refiners that adds quality and detail, and says it can also be used for inpainting or upscaling.

Send the `image` to refine and a `prompt` that describes it. Add a `mask` to refine only part of the image, or set `resolution` to `1024` or `2048`. Each request returns a list with one refined image URL.

## Features
- **Prompt-guided refinement**: `prompt` and `negative_prompt` steer the added detail.
- **Masked refinement**: pass a `mask` (same size as the image) to refine only that section.
- **Output resolution**: `original` (default), `1024` or `2048`.
- **Control over change**: `creativity` (denoising strength, 0 to 1) and `resemblance` (ControlNet conditioning scale, 0 to 1) set how far the result moves from the input.
- **HDR**: `hdr` (0 to 1) adds an HDR improvement over the original.
- **Sampler settings**: `scheduler` (`DDIM`, `DPMSolverMultistep`, `K_EULER_ANCESTRAL`, `K_EULER`), `steps`, `guidance_scale` and `seed`.

## Use Cases
1. **Generated images**: add texture and detail to images from other image generators.
2. **Portraits and product shots**: sharpen detail in skin, fabric or material surfaces guided by a prompt.
3. **Fixing one area**: refine only a face, a hand or an object by sending a mask.
4. **Larger output**: refine an image and return it at `1024` or `2048` resolution.

## Visual Examples
### Example Input
Input image, refined with the prompt "UHD 4k vogue, a woman resting in a magic pool, face above the surface of the water, red freckles" at the original resolution:

![Example input image sent to Magic Image Refiner](https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-input-api-market.webp)

```json
{
  "hdr": 0,
  "image": "https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-input-api-market.webp",
  "steps": 20,
  "prompt": "UHD 4k vogue, a woman resting in a magic pool, face above the surface of the water, red freckles",
  "scheduler": "DDIM",
  "creativity": 0.25,
  "guess_mode": false,
  "resolution": "original",
  "resemblance": 0.75,
  "guidance_scale": 7,
  "negative_prompt": "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant"
}
```

### Example Output
![Magic Image Refiner output: the refined image of a woman resting in a pool with red freckles](https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-output-xeh3vg4t-api-market-1.jpg)

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | none |  | Prompt that describes the image |
| `image` | string (URL) | No | none |  | URL of the image to refine |
| `mask` | string (URL) | No | none |  | Optional mask image URL. When provided, only that section of the image is refined. Must be the same size as the image |
| `resolution` | string | No | `original` | `original`, `1024`, `2048` | Output image resolution. `original` keeps the input size |
| `resemblance` | number | No | `0.75` | `0` to `1` | Conditioning scale for ControlNet (how closely the result follows the input image) |
| `creativity` | number | No | `0.25` | `0` to `1` | Denoising strength. 1 means the original image is fully replaced |
| `hdr` | number | No | `0` | `0` to `1` | HDR improvement over the original image |
| `scheduler` | string | No | `DDIM` | `DDIM`, `DPMSolverMultistep`, `K_EULER_ANCESTRAL`, `K_EULER` | Sampling scheduler |
| `steps` | integer | No | `20` |  | Number of denoising steps |
| `guidance_scale` | number | No | `7` | `0.1` to `30` | Scale for classifier-free guidance |
| `seed` | integer | No | none |  | Random seed for reproducible results |
| `negative_prompt` | string | No | `teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant` |  | Things to avoid in the result |
| `guess_mode` | boolean | No | `false` |  | In this mode the ControlNet encoder tries to recognize the content of the input image even if you remove all prompts. A `guidance_scale` between 3.0 and 5.0 is recommended |

No field is marked as required, but a request needs an `image` to refine.

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.135 per API unit.
- **1 API unit = 1 request with up to 124 seconds of processing.** A request that takes longer uses one unit per started block of 124 seconds (for example, 130 seconds of processing uses 2 units, $0.27).
- **Example cost**: the example took 3.8 seconds of processing, which is 1 unit, $0.135.
- **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/magic-image-refinement-upscaling-fermatresearch-magic-image-refiner`

### 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/magic-image-refinement-upscaling-fermatresearch-magic-image-refiner/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "507ddf6f977a7e30e46c0daefd30de7d563c72322f9e4cf7cbac52ef0f667b13",
    "input": {
      "hdr": 0,
      "image": "https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-input-api-market.webp",
      "steps": 20,
      "prompt": "UHD 4k vogue, a woman resting in a magic pool, face above the surface of the water, red freckles",
      "scheduler": "DDIM",
      "creativity": 0.25,
      "guess_mode": false,
      "resolution": "original",
      "resemblance": 0.75,
      "guidance_scale": 7,
      "negative_prompt": "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant"
    }
  }'
```

Response (201 Created):
```json
{
  "id": "xeh3vg4tz1rmw0d1ccn95dms3r",
  "version": "507ddf6f977a7e30e46c0daefd30de7d563c72322f9e4cf7cbac52ef0f667b13",
  "status": "starting",
  "created_at": "2026-10-22T10:31:39.139Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/magic-image-refinement-upscaling-fermatresearch-magic-image-refiner/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "xeh3vg4tz1rmw0d1ccn95dms3r",
  "status": "succeeded",
  "output": [
    "https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-output-xeh3vg4t-api-market-1.jpg"
  ],
  "error": null,
  "created_at": "2026-10-22T10:31:39.139Z",
  "started_at": "2026-10-22T10:34:09.677Z",
  "completed_at": "2026-10-22T10:34:13.431Z"
}
```

`output` is a list of URLs. Each request returns one refined image.

### 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 `resolution` that is not in the list or a `creativity` above 1.
- **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/magic-image-refinement-upscaling-fermatresearch-magic-image-refiner";

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: "507ddf6f977a7e30e46c0daefd30de7d563c72322f9e4cf7cbac52ef0f667b13", 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({
  "hdr": 0,
  "image": "https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-input-api-market.webp",
  "steps": 20,
  "prompt": "UHD 4k vogue, a woman resting in a magic pool, face above the surface of the water, red freckles",
  "scheduler": "DDIM",
  "creativity": 0.25,
  "guess_mode": false,
  "resolution": "original",
  "resemblance": 0.75,
  "guidance_scale": 7,
  "negative_prompt": "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant"
}).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/magic-image-refinement-upscaling-fermatresearch-magic-image-refiner"


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": "507ddf6f977a7e30e46c0daefd30de7d563c72322f9e4cf7cbac52ef0f667b13", "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({
    "hdr": 0,
    "image": "https://blog.api.market/wp-content/uploads/2026/10/magic-image-refiner-input-api-market.webp",
    "steps": 20,
    "prompt": "UHD 4k vogue, a woman resting in a magic pool, face above the surface of the water, red freckles",
    "scheduler": "DDIM",
    "creativity": 0.25,
    "guess_mode": False,
    "resolution": "original",
    "resemblance": 0.75,
    "guidance_scale": 7,
    "negative_prompt": "teeth, tooth, open mouth, longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, mutant"
}))
```

## FAQ
- **How do I keep the result close to my image?**
  - Lower `creativity` (denoising strength) and raise `resemblance`. `creativity` set to 1 fully replaces the original image.
- **Can I refine only part of an image?**
  - Yes. Send a `mask` that is the same size as the image. Only that section is refined.
- **Can it make the image larger?**
  - Set `resolution` to `1024` or `2048`. The default `original` keeps the input size.
- **What does one request cost?**
  - $0.135 for a request with up to 124 seconds of processing. The example took 3.8 seconds and cost $0.135.
- **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.


Add detail and quality to an existing image with Magic Image Refiner, with optional masked refinement and 1024 or 2048 output resolution.