Models
Each model has a credit cost and an estimated provider cost, recorded on the job. Routes fall back in order. Runware and Black Forest Labs sit in the registry, switched off until they are connected.
FLUX Schnell
The fast draft model. It is the only one the daily free credits can buy.
1 credits
Use this modelFLUX Dev
A slower FLUX model for stills that need more care than a draft.
3 credits
Use this modelFLUX Pro 1.1
FLUX Pro 1.1 for campaign stills, jewelry, and packaging you might actually print.
5 credits
Use this modelFLUX Pro 1.1 Ultra
FLUX Pro 1.1 Ultra renders up to four megapixels with a natural, photographic finish.
3 credits
Use this modelSeedream 4.5
ByteDance's Seedream 4.5 makes sharp, detailed stills at about four megapixels.
2 credits
Use this modelNano Banana 2.1 (Google)
Google's Nano Banana 2.1 follows long prompts closely and keeps scenes coherent.
2 credits
Use this modelGPT Image 2
OpenAI's GPT Image 2 is strong at legible text and instruction-following.
3 credits
Use this modelGrok Imagine
xAI's Grok Imagine makes quick, stylish stills at a low price.
1 credits
Use this modelRecraft V4
Recraft V4 is built for design work: clean layouts, brand-ready styles, and vector-like art.
2 credits
Use this modelIdeogram V3
Ideogram V3 handles typography and graphic composition better than most models.
3 credits
Use this modelFLUX Kontext Pro
FLUX Kontext edits a photo you upload from a short written instruction while keeping the rest intact.
2 credits
Use this modelSeedream 4.5 Edit
Seedream 4.5 Edit reworks an uploaded image from a text instruction at high resolution.
2 credits
Use this modelGPT Image 2 Edit
GPT Image 2 Edit changes an uploaded image from a written instruction and follows it closely.
3 credits
Use this modelClarity Upscaler
Clarity Upscaler enlarges an image and adds believable detail.
6 credits
Use this modelReal-ESRGAN Upscaler
Real-ESRGAN is a fast, low-cost upscaler that stays faithful to the source.
1 credits
Use this modelBria Background Removal
Bria removes the background from an uploaded image and returns a transparent PNG.
1 credits
Use this modelKling 4.0
Kling 4.0 makes smooth, cinematic video from text or a starting image.
from 36 credits
Use this modelVeo 3.1 Fast
Google's Veo 3.1 Fast is the quicker, cheaper Veo, with optional generated sound.
from 19 credits
Use this modelVeo 3.1
Google's Veo 3.1 gives the highest Veo quality, with optional generated sound.
from 37 credits
Use this modelMiniMax Hailuo H3
MiniMax Hailuo H3 makes expressive video, from 480p up to 2K.
from 14 credits
Use this modelWan 3.0
Wan 3.0 is a capable video model with a low entry price.
from 23 credits
Use this modelSeedance 2.5
ByteDance's Seedance 2.5 is the newest Seedance, with strong motion and prompt adherence.
from 108 credits
Use this modelSeedance 2.5 Reference
Seedance 2.5 Reference animates a clip guided by your uploaded image. Refer to it as @Image1 in the prompt.
from 108 credits
Use this modelSeedance 2.0
Seedance 2.0 is ByteDance's cinematic video model, with native audio and multi-shot scenes.
from 70 credits
Use this modelSeedance 2.0 Fast
Seedance 2.0 Fast is the lower-cost tier of Seedance 2.0, up to 720p.
from 56 credits
Use this modelSeedance 1.0 Lite
Seedance 1.0 Lite is the light first-generation Seedance. The provider now routes it to Seedance 1.0 Pro Fast.
from 5 credits
Use this modelGrok Imagine Video
xAI's Grok Imagine Video makes short clips from text or an image, from 1 second.
from 20 credits
Use this model