flux-schnell

Generate a 1 megapixel image from a text prompt in roughly one second on a warm worker. The model is Black Forest Labs' four-step diffusion checkpoint, hosted on Replicate and called through the Socaity backend, so a single Socaity API key covers everything. The default output is WebP at quality 80. PNG and JPEG are also supported.

by black-forest-labs · socaity.sdk.replicate.black_forest_labs.flux_schnell

Capabilities

Text-to-image generation in four denoising steps (around one second per image on a warm worker)
Throughput-tuned mode (go_fast=True, the default); set go_fast=False for the higher-precision path
Configurable aspect ratio: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, and more
Output format webp (default), png, or jpg, with quality 0 to 100
Up to four images per call via num_outputs
Reproducible generation with an integer seed
Built-in safety checker, with opt-out via disable_safety_checker

Install

pip install socaity
socaity install black-forest-labs/flux-schnell

Authentication

Socaity proxies the Replicate call for you, so one API key covers everything. Set SOCAITY_API_KEY in your environment. You do not need a Replicate key. Get a Socaity key at socaity.ai/account/secrets.

export SOCAITY_API_KEY="sk_..."

Quick start

Generate one 1 MP WebP image at the default settings. .get_result() blocks until the job finishes and returns one image object per requested output.

import os
from socaity.sdk.replicate.black_forest_labs import flux_schnell

flux = flux_schnell(api_key=os.getenv("SOCAITY_API_KEY"))

# predictions(...) submits the job and returns a Job handle.
# flux(prompt=...) and flux.run(prompt=...) are aliases for the same call.
image = flux.predictions(
    prompt="a cat sitting on mars, cinematic lighting",
    go_fast=True,
    megapixels="1",
    num_outputs=1,
    aspect_ratio="1:1",
    output_format="webp",
    output_quality=80,
    num_inference_steps=4,
).get_result()

image.save("output.webp")

Methods

predictions(prompt, seed=42, go_fast=True, megapixels="1", num_outputs=1, aspect_ratio="1:1", output_format="webp", output_quality=80, num_inference_steps=4, disable_safety_checker=False, **kwargs) -> Job

Submit a text-to-image job and return a Job object. flux(prompt=...) and flux.run(prompt=...) are aliases that call the same endpoint.

Parameter	Type	Default	Description
`prompt`	`str`	`required`	Text description of the image to generate.
`seed`	`int`	`42`	Seed for reproducible generation. Defaults to a fixed value of 42 for deterministic output, not a random seed; change it to vary results.
`go_fast`	`bool`	`True`	Throughput-tuned generation path. Set to False for the higher-precision path.
`megapixels`	`str`	`"1"`	Approximate output resolution in megapixels.
`num_outputs`	`int`	`1`	Number of images to generate per call.
`aspect_ratio`	`str`	`"1:1"`	Aspect ratio of the output image (e.g. "16:9", "9:16", "4:3").
`output_format`	`str`	`"webp"`	Output image format. Options: webp, png, jpg.
`output_quality`	`int`	`80`	Compression quality for webp and jpg outputs (0 to 100). Not applied to png.
`num_inference_steps`	`int`	`4`	Denoising steps. 4 is the recommended value; lower values reduce quality.
`disable_safety_checker`	`bool`	`False`	Disable the built-in safety filter on generated images.

Pricing & live playground

Current pricing and the interactive playground live on socaity.ai.

Open flux-schnell on socaity.ai

Next steps

Run your first model: end-to-end walkthrough of the install, auth, and first job.
Multi-model pipeline: chain flux_schnell with a hosted text-to-speech model to produce an image and a narration in one script.
Python SDK reference: full surface of FastSDK, job lifecycle, and gather helpers.

Audio-Driven Face Video

deepseek-v3