Name: FormFinder FLUX API: Moody Architecture Image Generator
Brand: magicapi
SKU: api-magicapi-moody-atmospheric-image-generator-ismail-seleit-formfinder-flux
Availability: InStock
Author: API Market

Shashank Agarwal

# 🌌 Atmospheric Architecture Image Generator API (ismail-seleit FormFinder FLUX)

## Overview
FormFinder FLUX is a LoRA style for the FLUX image model, published by ismail-seleit as the FLUX version of FormFinder-XL. Its developer describes it as trained to create moody atmospheric images, and versatile enough to mix with other LoRAs for diverse styles with interesting architectural shapes.

Send a text `prompt` and the model returns a list of image URLs (1 to 4 images, set by `num_outputs`). You can also pass an `image` for image-to-image, or an `image` plus a `mask` for inpainting.

## Features
- **Moody, atmospheric style** with unusual architectural forms.
- **Mix with another LoRA**: load a second style with `extra_lora` and balance the two with `lora_scale` and `extra_lora_scale`.
- **Text to image, image to image and inpainting**: use `prompt` alone, add `image` (with `prompt_strength`), or add `image` and `mask`.
- **Aspect ratios**: `1:1` (default), `16:9`, `21:9`, `3:2`, `2:3`, `4:5`, `5:4`, `3:4`, `4:3`, `9:16`, `9:21`, or `custom` with `width` and `height` (256 to 1440 px).
- **Two base models**: `dev` (default, about 28 steps) or `schnell` (about 4 steps).
- **Up to 4 images per request** with `num_outputs`.
- **Output formats**: `webp` (default), `jpg` or `png`, with `output_quality` 0 to 100.
- **Speed mode**: `go_fast` runs an fp8-quantized model for faster results.

## Use Cases
1. **Architectural concept art**: speculative buildings and structures in dramatic light.
2. **Mood boards**: atmospheric scenes for film, game or event concepts.
3. **Style blends**: combine this style with a second LoRA for new looks.
4. **Restyling photos**: send a photo of a building or street as `image` and reinterpret it.

## Visual Examples
### Example Input
Prompt (shortened): "Envision an otherworldly skyscraper suspended mid-air, defying physics with an unprecedented, avant-garde design that blurs architecture and art. ... Situated amidst the bustling streets of central Tokyo ... Frame the scene with an antique Rolleiflex TLR camera ..." The full prompt is in the request below.

Settings: `model` `schnell`, `num_inference_steps` `4`, `guidance_scale` `1`, `aspect_ratio` `1:1`, `width` `1024`, `height` `1024`, `lora_scale` `0.63`, `extra_lora_scale` `0.8`, `num_outputs` `1`, `output_format` `webp`, `output_quality` `80`. Because `aspect_ratio` is not `custom`, `width` and `height` have no effect here, and no `extra_lora` is loaded, so `extra_lora_scale` is unused.

### Example Output
![FormFinder FLUX output for the floating Tokyo skyscraper prompt](https://blog.api.market/wp-content/uploads/2026/10/formfinder-flux-output-c5gfn1zb-api-market.webp)

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | Yes |  |  | Prompt for the generated image |
| `image` | string (URL) | No | none |  | Input image for image-to-image or inpainting mode. If provided, `aspect_ratio`, `width` and `height` are ignored |
| `mask` | string (URL) | No | none |  | Image mask for inpainting mode. If provided, `aspect_ratio`, `width` and `height` are ignored |
| `aspect_ratio` | string | No | `1:1` | `1:1`, `16:9`, `21:9`, `3:2`, `2:3`, `4:5`, `5:4`, `3:4`, `4:3`, `9:16`, `9:21`, `custom` | Aspect ratio for the generated image. `custom` uses `width` and `height` and runs in bf16 mode |
| `height` | integer | No | none | 256 to 1440 | Height of the generated image. Only works if `aspect_ratio` is `custom`. Rounded to the nearest multiple of 16. Incompatible with `go_fast` |
| `width` | integer | No | none | 256 to 1440 | Width of the generated image. Only works if `aspect_ratio` is `custom`. Rounded to the nearest multiple of 16. Incompatible with `go_fast` |
| `prompt_strength` | number | No | `0.8` | 0 to 1 | Prompt strength when using image-to-image. 1.0 fully replaces the information in the input image |
| `model` | string | No | `dev` | `dev`, `schnell` | Which model to run. `dev` performs best with around 28 inference steps; `schnell` only needs 4 |
| `num_outputs` | integer | No | `1` | 1 to 4 | Number of images to generate |
| `num_inference_steps` | integer | No | `28` | 1 to 50 | Number of denoising steps. More steps can give more detailed images, but take longer |
| `guidance_scale` | number | No | `3` | 0 to 10 | Guidance scale for the diffusion process. Lower values can give more realistic images. Good values to try are 2, 2.5, 3 and 3.5 |
| `seed` | integer | No | random |  | Random seed. Set it for reproducible generation |
| `output_format` | string | No | `webp` | `webp`, `jpg`, `png` | Format of the output images |
| `output_quality` | integer | No | `80` | 0 to 100 | Quality when saving the output images. 100 is best quality. Not relevant for `png` |
| `disable_safety_checker` | boolean | No | `false` |  | Disable the safety checker for generated images |
| `go_fast` | boolean | No | `false` |  | Run faster predictions with a model optimized for speed (currently fp8 quantized); disable to run in original bf16 |
| `megapixels` | string | No | `1` | `1`, `0.25` | Approximate number of megapixels for the generated image |
| `lora_scale` | number | No | `1` | -1 to 3 | How strongly the main LoRA is applied. Sane results between 0 and 1 for base inference. With `go_fast` a 1.5x multiplier is applied to this value |
| `extra_lora` | string | No | none |  | Extra LoRA weights to load: a Hugging Face URL (`huggingface.co/<owner>/<model-name>`), a CivitAI URL (`civitai.com/models/<id>[/<model-name>]`) or a direct `.safetensors` URL |
| `extra_lora_scale` | number | No | `1` | -1 to 3 | How strongly the extra LoRA is applied. Sane results between 0 and 1 for base inference. With `go_fast` a 1.5x multiplier is applied to this value |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.0034 per API unit.
- **1 API unit = 1 successful request with up to 2 seconds of processing.** A longer job bills one more unit for each started block of 2 seconds (for example, 7 seconds of processing is 4 units, $0.0136).
- Processing time grows with `num_outputs`, `num_inference_steps` and image size; `schnell` with 4 steps is the fastest setting.
- **Example cost**: the example request (`schnell`, 4 steps, one image) took 2.4 seconds of processing, which is 2 units, so $0.0068. The default `dev` model with 28 steps takes longer and bills more units.
- **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/moody-atmospheric-image-generator-ismail-seleit-formfinder-flux`

### 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/moody-atmospheric-image-generator-ismail-seleit-formfinder-flux/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "58c5c83ee1d1e9ff051b647b459d737adacbc0194d502236d66e3b25f4affe9a",
    "input": {
      "model": "schnell",
      "width": 1024,
      "height": 1024,
      "prompt": "Envision an otherworldly skyscraper suspended mid-air, defying physics with an unprecedented, avant-garde design that blurs architecture and art. This gravity-defying marvel boasts undulating curves, twisted steel beams, and glowing, iridescent surfaces that seem to shift and shimmer in the light. Situated amidst the bustling streets of central Tokyo, this impossible structure pierces the sky like a shard of futuristic promise, its proximity to towering skyscrapers, neon-lit billboards, and crowds of pedestrians creating a jarring yet fascinating juxtaposition between the old and new, the organic and synthetic.\n\nFrame the scene with an antique Rolleiflex TLR camera, its twin-lens design and leaf shutter evoking a sense of nostalgia and timelessness. Employ a deliberate, artistic approach to composition, emphasizing the play of shadows across the building'\''s surface, the reflections off its metallic skin, and the ethereal mist that shrouds its base.\n\nThe camera captures the symphony of sounds - the hum of traffic, the chatter of pedestrians, the wail of sirens - as a gentle, pulsing hum that underscores the tension between the impossible and the mundane. As the shutter clicks shut, the world around this fantastical building seems to fade into a soft, diffused glow, while the structure itself remains sharp, crystalline, and resolute in its defiance of physics.",
      "lora_scale": 0.63,
      "num_outputs": 1,
      "aspect_ratio": "1:1",
      "output_format": "webp",
      "guidance_scale": 1,
      "output_quality": 80,
      "extra_lora_scale": 0.8,
      "num_inference_steps": 4
    }
  }'
```

Response (201 Created):
```json
{
  "id": "c5gfn1zbz5rmr0d1d8gsh88atm",
  "version": "58c5c83ee1d1e9ff051b647b459d737adacbc0194d502236d66e3b25f4affe9a",
  "status": "starting",
  "created_at": "2026-10-23T18:59:30.711Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/moody-atmospheric-image-generator-ismail-seleit-formfinder-flux/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "c5gfn1zbz5rmr0d1d8gsh88atm",
  "status": "succeeded",
  "output": [
    "https://blog.api.market/wp-content/uploads/2026/10/formfinder-flux-output-c5gfn1zb-api-market.webp"
  ],
  "error": null,
  "created_at": "2026-10-23T18:59:30.711Z",
  "started_at": "2026-10-23T18:59:31.711Z",
  "completed_at": "2026-10-23T18:59:33.711Z"
}
```

`output` is an array of image URLs, one per generated 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 missing `prompt`, a `num_outputs` above 4 or an `aspect_ratio` 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/moody-atmospheric-image-generator-ismail-seleit-formfinder-flux";

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: "58c5c83ee1d1e9ff051b647b459d737adacbc0194d502236d66e3b25f4affe9a", 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({
  "model": "schnell",
  "width": 1024,
  "height": 1024,
  "prompt": "Envision an otherworldly skyscraper suspended mid-air, defying physics with an unprecedented, avant-garde design that blurs architecture and art. This gravity-defying marvel boasts undulating curves, twisted steel beams, and glowing, iridescent surfaces that seem to shift and shimmer in the light. Situated amidst the bustling streets of central Tokyo, this impossible structure pierces the sky like a shard of futuristic promise, its proximity to towering skyscrapers, neon-lit billboards, and crowds of pedestrians creating a jarring yet fascinating juxtaposition between the old and new, the organic and synthetic.\n\nFrame the scene with an antique Rolleiflex TLR camera, its twin-lens design and leaf shutter evoking a sense of nostalgia and timelessness. Employ a deliberate, artistic approach to composition, emphasizing the play of shadows across the building's surface, the reflections off its metallic skin, and the ethereal mist that shrouds its base.\n\nThe camera captures the symphony of sounds - the hum of traffic, the chatter of pedestrians, the wail of sirens - as a gentle, pulsing hum that underscores the tension between the impossible and the mundane. As the shutter clicks shut, the world around this fantastical building seems to fade into a soft, diffused glow, while the structure itself remains sharp, crystalline, and resolute in its defiance of physics.",
  "lora_scale": 0.63,
  "num_outputs": 1,
  "aspect_ratio": "1:1",
  "output_format": "webp",
  "guidance_scale": 1,
  "output_quality": 80,
  "extra_lora_scale": 0.8,
  "num_inference_steps": 4
}).then((urls) => console.log(urls));
```

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

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/moody-atmospheric-image-generator-ismail-seleit-formfinder-flux"


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": "58c5c83ee1d1e9ff051b647b459d737adacbc0194d502236d66e3b25f4affe9a", "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({
    "model": "schnell",
    "width": 1024,
    "height": 1024,
    "prompt": "Envision an otherworldly skyscraper suspended mid-air, defying physics with an unprecedented, avant-garde design that blurs architecture and art. This gravity-defying marvel boasts undulating curves, twisted steel beams, and glowing, iridescent surfaces that seem to shift and shimmer in the light. Situated amidst the bustling streets of central Tokyo, this impossible structure pierces the sky like a shard of futuristic promise, its proximity to towering skyscrapers, neon-lit billboards, and crowds of pedestrians creating a jarring yet fascinating juxtaposition between the old and new, the organic and synthetic.\n\nFrame the scene with an antique Rolleiflex TLR camera, its twin-lens design and leaf shutter evoking a sense of nostalgia and timelessness. Employ a deliberate, artistic approach to composition, emphasizing the play of shadows across the building's surface, the reflections off its metallic skin, and the ethereal mist that shrouds its base.\n\nThe camera captures the symphony of sounds - the hum of traffic, the chatter of pedestrians, the wail of sirens - as a gentle, pulsing hum that underscores the tension between the impossible and the mundane. As the shutter clicks shut, the world around this fantastical building seems to fade into a soft, diffused glow, while the structure itself remains sharp, crystalline, and resolute in its defiance of physics.",
    "lora_scale": 0.63,
    "num_outputs": 1,
    "aspect_ratio": "1:1",
    "output_format": "webp",
    "guidance_scale": 1,
    "output_quality": 80,
    "extra_lora_scale": 0.8,
    "num_inference_steps": 4
}))
```

## FAQ
- **Is there a trigger word?**
  - No trigger word is documented for this style. Describe the scene, architecture and mood you want in `prompt`.
- **Is my whole prompt used?**
  - Not always. Very long prompts are truncated: in a run of the example request the model logged that the CLIP text encoder keeps the first 77 tokens and the second text encoder keeps the first 256. Put the most important details first.
- **How many images does one request return?**
  - Between 1 and 4, set by `num_outputs` (default 1). Billing is by processing time, so more images take longer and can add units.
- **How do I combine it with another style?**
  - Set `extra_lora` to a Hugging Face, CivitAI or `.safetensors` LoRA URL and balance the two with `lora_scale` and `extra_lora_scale`.
- **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.


Create moody, atmospheric images and speculative architecture with the FormFinder FLUX LoRA: text-to-image, image-to-image and inpainting.