Getting started

Sumi Studio is an AI creation tool. You can start a conversation with the assistant on the home page, or open a canvas and wire image, video, audio and text nodes together to generate work.

This handbook matches the live interface. Pick a chapter on the left; each page is one chapter. Work through this chapter once and you will go from signing up to your first creation.

Sign up and sign in

• Click Sign in at the top right. You can sign in with a password or a verification code (password is the default); Google / Discord / Telegram (shown per site config) and wallet-signature sign-in are supported too.
• While signed out, anything that needs an account (new canvas, copy project, publish, and so on) opens the sign-in dialog instead of sending you away. Only the canvas editor itself requires you to sign in first.
• Starter credits, if any, follow the sign-up page and the credit ledger. Check the live model quote before you generate.

First creation: start from the home assistant

The home page centers on the assistant's input box:

• Use the Create / Film & drama / Marketing tabs above the box to switch among the three assistants;
• Say what you want in the box. Add attachment uploads files or adds them from your library, @ references media, Skills lists the skills you can use, and the Canvas switch opens the work on a canvas;
• Pick a generation mode before you send: in Manual mode every generation shows its price and waits for your confirmation; in Auto mode the assistant places generations without asking each time, still charged at the quoted price and limited by your balance and budget.

From other pages, click Create at the top of the left sidebar or press ⌘/Ctrl + K to come back to this box.

You can also open a canvas directly

• Click the + next to Recent projects in the left sidebar and choose New project or New canvas; Projects (/projects) lists all your projects and can create them too;
• Open a template in the template library (/templates) and click Duplicate canvas, or click Use template on the Templates tab under Profile in the account menu. Either way you get an editable canvas based on the template.

Your first text-to-image on an empty canvas is four steps: double-click empty space and pick Image; type the picture in the generator bar below the node; confirm the credit estimate; click send or press ⌘/Ctrl + Enter. The result returns to that node. Edits save automatically.

The pieces of the interface to know

Once signed in, the left sidebar is the main navigation on desktop. The right side of the top bar holds the appearance toggle, language, your credit balance (click it for the Credit Store), membership and the account menu.

• Top of the sidebar: Create (⌘/Ctrl + K) takes you back to the Sumi Harness box on the home page;
• Below it: Projects, Asset Library, Skills, Community Works and MCP;
• Then Recent projects, where the + next to the heading creates a project or canvas; Help & feedback (Guides, Contact us) sits at the bottom;
• Account menu: membership and storage cards at the top (subscribe / upgrade, redeem code), then Profile, Subscription, Usage, AI watermark, Tutorials & Docs, Messages, language, Appearance, Settings, Log out and more. On mobile, the menu button at the top left opens the sidebar, with appearance and language at its bottom.