Quick start

From nothing to a first picture made in ComfyUI from a graph you built and understand. Plan an hour the first time, most of it for installing ComfyUI and downloading a model.

1. Get ComfyUI running

Install ComfyUI, put one checkpoint model into ComfyUI/models/checkpoints, start ComfyUI and open its address in the browser (usually http://127.0.0.1:8188). Details: ComfyUI and where to get it.

2. Install NodeWise Studio

sudo apt install ./nodewise-studio_*_amd64.deb

Get the file from the project's Releases page. Details: Installing.

3. Start it

Open "NodeWise Studio" from the menu. It speaks your system's language; the button in the top corner of the welcome card switches between Русский and English. The tour of the window runs once; Skip the tour if you like. The welcome card offers three ways to start (Take a lesson, Open an example, Ask the assistant) and, under them, what they need: Connect ComfyUI and Choose the assistant. No ComfyUI yet? Choose I have no ComfyUI yet. See The welcome card.

4. Connect ComfyUI and read it

Choose Connect… on the welcome card. In Settings ▸ Connection choose Find on this computer, or enter the address you open ComfyUI at in the browser (for example http://127.0.0.1:8188), choose Check ComfyUI, then Save settings. Back on the card choose Read ComfyUI. When it is done, the card says how many node types and model files it read, and the library on the left shows them. The whole chapter: Connecting to ComfyUI.

5. Take a first lesson

Choose Take a lesson on the welcome card (or choose Learn in the right panel). Pick the lesson A first picture from text and follow its steps. You add the nodes and links yourself; the lesson checks each step on the graph and says in words what is missing. Show me where points at the place on the canvas; Do this for me does a step where the choice is not yours to make. No assistant model is needed. See A first picture from text.

6. Check and save for ComfyUI

  1. Graph ▸ Check graph (Ctrl+Shift+C). Fix what it finds, or read why a finding is fine.
  2. File ▸ Save workflow for ComfyUI… (Ctrl+E) and choose where to put the .json file.

7. Run it in ComfyUI

After saving, the card Run it in ComfyUI opens. Choose Open ComfyUI, drag the .json file onto the ComfyUI page in the browser (or Copy workflow and press Ctrl+V on the ComfyUI canvas), and press Run there. The picture appears in the Save Image node and is written to ComfyUI/output. Details: Running the result in ComfyUI.

8. Optional: an assistant

To ask questions about any node or let the assistant build, choose who answers: a cloud model with your key, or a model on your own computer (Ollama, LM Studio, vLLM, the llama.cpp server). Choose the assistant on the welcome card or in Settings ▸ Assistant. See Choosing who answers. Cloud answers cost money per question.

See also

This page is imported from the application repository's wiki (docs/wiki/start/Quick-Start.md, revision b8e3f71). The same guide is built into the app: Help ▸ User guide.

To refresh it, run npm run import and then npm run build in the website folder.