Pay for finished tasks, not for setting up ComfyUI.
Bring your ComfyUI workflow, and Studio runs it on a GPU that’s ready for it. We pick the right GPU and set up the nodes and models for you.
API usersRun your own workflows from your code, from n8n or from an AI agent.Upload your workflow's .json, and Studio gets it ready to run.
Studio sets up the nodes and models your workflow needs and picks the right GPU for it. Your graph becomes a form anyone on your team can fill in.
Missing a model? Paste its Hugging Face or Civitai link.
Studio picks the right GPU for each workflow.
You never choose hardware. Studio matches each workflow to a GPU that fits it and moves to a bigger one if a task needs it.
Run one workflow on a hundred files, or chain several together.
Pick files in Media and run them all at once. Before anything starts, you can skip combinations you don’t want and swap files.
Each file you pick gets its own run.
Run every combination, or just enough runs to use each file at least once.
Chain steps such as edit, video and upscale, on one file or a hundred.
Send a result to another workflow as its input, without uploading it again.
Rerun failed tasks in one click.
Results are saved together, each linked to the files it came from.
Ready-made apps if you don't have a workflow.
Video
A short film of several shots, with its own sound.
Photo
A new photo from a prompt and a few pictures.
Photo edit
Edit a photo with a prompt, one change at a time.
Video and Photo are available in most countries.
Run your workflows from your code or your AI agent.
You keep your own LoRAs, graphs and chains.
Use every setting the model offers, with nothing cut down.
Video, Photo and Photo edit are as easy to use as an API call.
Call Studio from your own code or from n8n. It runs your workflows at the same prices as the app.
Ask Claude or ChatGPT to run your workflows. They connect to Studio over MCP with your account.
POST /v1/studio/runs { "workflow": "<your workflow>", "inputs": { "Photo": 8412, "Prompt": "make the jacket red" }, "webhook_url": "https://your.app/hooks/studio" } # when it's done → run.completed { "status": "completed", "outputs": [{ "media_id": 9133, "kind": "image" }] }
What to know before your first task.
Will my custom nodes work?
Most do. Studio installs the packs your workflow uses and tells you before your first run if one can’t work.
What about my models and LoRAs?
Paste a Hugging Face or Civitai link. For gated models, add your own Hugging Face key.
Will a heavy video graph fit?
Usually. Studio picks a GPU big enough and moves to a bigger one if a task needs it. If a workflow can’t run, it tells you.
How much does a task cost?
You pay per task, by how long it runs. Every task and its price are in your billing history.
Is there an API?
Yes. Run your workflows from your code, from n8n or from an AI agent, at the same prices. The apps are there when you’d rather click.
Do I need a subscription?
No. Top up your balance and pay for each task.
Where do my files go?
To Media, your file library.