ReplicateImageComing soon

black-forest-labs/flux-fill-dev

black-forest-labs/flux-fill-dev is an image tool from Replicate that runs in the background and returns images. Coming soon: listed with its price, not yet open to runs.

This tool is listed but not yet available to run.

Pricing

Coming soon

$0.052 per image. Billed on num_outputs, default 1.

Limits

Runs
In the background
Returns
image
Longest queue wait
1 hour
Run deadline
30 minutes
Longest run, end to end
225 minutes
Largest inline input
1 MB
Largest input file
500 MB
Largest output
5 GB
Outputs per run
16

Input

FieldKindDefaultDescription
image requiredURL or file—

The image to inpaint. Can contain alpha mask. If the image width or height are not multiples of 32, they will be scaled to the closest multiple of 32. If the image dimensions don't fit within 1440x1440, it will be scaled down to fit.

prompt requiredtext—

Prompt for generated image

masktext—

A black-and-white image that describes the part of the image to inpaint. Black areas will be preserved while white areas will be inpainted.

seednumber (whole)—

Random seed. Set for reproducible generation

guidancenumber
0 to 100
30

Guidance for generated image

lora_scalenumber
-1 to 3
1

Determines how strongly the main LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.

megapixelsJSON1

Approximate number of megapixels for generated image. Use match_input to match the size of the input (with an upper limit of 1440x1440 pixels)

num_outputsnumber (whole)
1 to 4
1

Number of outputs to generate

lora_weightstext—

Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'

output_formatJSONwebp

Format of the output images

output_qualitynumber (whole)
0 to 100
80

Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs

num_inference_stepsnumber (whole)
1 to 50
28

Number of denoising steps. Recommended range is 28-50, and lower number of steps produce lower quality outputs, faster.

Same request in code

curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "replicate",
    "model": "black-forest-labs/flux-fill-dev",
    "input": {
      "image": "<image>",
      "prompt": "<prompt>"
    }
  }'