recipe
412
tiny film, 12 seconds
@nico
fluxkling-2.5suno
Open and fork
An infinite AI-powered canvas for creators
Make anything.
An infinite canvas for AI. Drop in any model, connect outputs to inputs, and build the thing only you would build.
01 — Every model
Every model.
One ⌘K away.
Drop in Flux, Kling, Suno, Veo — anything new this week. Press ⌘K, search, drag. The model is a node. Connect it to anything else. That’s the whole onboarding.
flux|↵
flux-1.1-proimage generation
kling-2.5image → video
veo-3text → video
minimax-musicsong generation
suno-v4music + lyrics
trellisimage → 3D
whisper-v3audio → text
sdxl-turbofast image
claude-sonnet-4.5reasoning
stable-audio-2sound effects
hunyuan-videotext → video
real-esrganupscale
elevenlabs-ttstext → speech
instant-meshimage → mesh
flux-1.1-proimage generation
kling-2.5image → video
veo-3text → video
minimax-musicsong generation
suno-v4music + lyrics
trellisimage → 3D
whisper-v3audio → text
sdxl-turbofast image
claude-sonnet-4.5reasoning
stable-audio-2sound effects
hunyuan-videotext → video
real-esrganupscale
elevenlabs-ttstext → speech
instant-meshimage → mesh
02 — Compose
Chain your creativity
together.
An idea becomes an image. The image becomes a film. The film picks up a score. Outputs feed inputs and the graph runs itself — branches in parallel, types matched, one canvas the whole way.
01 — text
An idea
A prompt. A reference. The smallest possible spark.
02 — image
An image
Generate a wide shot in flux-1.1-pro. Ten variants. Pick one.
03 — video
A film
Animate the still in kling-2.5 with a slow push-in.
04 — audio
A score
Score it in suno-v4. Twelve seconds. It fits.
03 — Recipes
The good ideas
spread.
Publish a workflow as a recipe. Send the link. Anyone can fork it, swap models, retune prompts, make it theirs. You watch what other people build, and you build on it.
recipe
1,284
looping product hero
@deia
sdxl-turboreal-esrgan
Open and fork
recipe
87
song from a journal entry
@maru
claudesuno-v4
Open and fork
And the rest of it
Train your own
LoRAs and custom fine-tunes, in the same canvas.
Parallel branches
The graph runs branches concurrently, automatically.
Smart node UI
Big models with 30 inputs collapse to the ones that matter.
⌘K everything
Find any model, any time. Search and drag.
Bring your media
Drag-drop images, video, audio, 3D files in.
Yours by default
Workflows are private until you publish a recipe.
Made with Infinite
What people are
already making.
12 workflows
refreshed daily
no asset · seed me
video
tiny film, 12 seconds
flux·kling-2.5·suno
no asset · seed me
image
looping product hero
sdxl-turbo·real-esrgan
no asset · seed me
audio
song from a journal entry
claude·suno-v4
no asset · seed me
image
an actor I made up
flux-pro
no asset · seed me
3d
spinning teapot
flux·trellis
no asset · seed me
video
rain at the corner store
flux·kling-2.5
no asset · seed me
image
type specimen, set in nothing
flux
no asset · seed me
video
a fox doing nothing
veo-3
no asset · seed me
audio
a 90s opening sting
minimax-music
no asset · seed me
image
the hat in 14 ways
sdxl-turbo·lora:fashion
no asset · seed me
3d
logo, but solid
instant-mesh
no asset · seed me
video
an alley that doesn't exist
flux·kling-2.5·elevenlabs
Pricing
Pay for what you run.
That’s it.
Free credits when you sign up. Top up whenever. No subscription, no seats, no minimums.
Starter
$10/ once
1,000 credits
Feel the canvas out. A first short film. A handful of stills.
Maker
$45/ once
5,000 credits
For people in the loop daily. Iteration room. Real projects.
Pro
$199/ once
25,000 credits
When you've stopped counting runs. Ship the album, the campaign.
Every plan includes
- every model
- training
- recipes
- private by default
- output ownership
- no expiry
$1 ≈ 100 credits · credits never expire
FAQ
Things you’ll wonder
before signing up.
- 01
How does pricing work?
Credits. Each model run costs credits proportional to its compute cost. Free credits when you sign up. Top up when you want. No subscription, no seats.
- 02
Which models are supported?
Every major image, video, audio, 3D, and language model — and the new ones show up automatically. If it ships, it's a node.
- 03
Do I own my outputs?
Yes. Everything you generate is yours. Files live in your account; download them anytime.
- 04
Can I train my own models?
Yes. LoRAs and custom fine-tunes run inside the canvas like any other node — same drag, same connect.
- 05
Can I export the workflow?
You can publish it as a recipe and share a link. Anyone you send it to can fork it. JSON export is on the roadmap.
- 06
Is my work private?
Yes, by default. Workflows are visible only to you until you publish them as a recipe.
- 07
Where do my files live?
In your account. Outputs are downloaded and re-hosted on our storage so model-provider links don't rot.