Introduction

Getting Started

By the end of this chapter, you should have Sweet Tea Studio installed, connected to ComfyUI, and through your first successful AI image. Once that baseline works, you can branch into variations or a short video pass without changing the basic mental model.

Sweet Tea is not here to make you stare at machinery. It is here to help you move from idea to result: choose a Pipe, change the controls that matter, generate, inspect the output, and keep the context that got you there.

What You'll Need

Keep the first run simple. Before you launch Sweet Tea for the first time, make sure you have:

• A supported machine — Public downloads are currently available for Windows and Linux.
• Disk space — Plan for at least 15 GB free for the app, one model, and your first outputs.
• A ComfyUI path — Either an existing install or a willingness to let Sweet Tea manage one for you.
• At least one model — A working checkpoint available to that ComfyUI engine.

Tip: You do not need a web account to generate locally. Account linking can come later if you want sync, Desk continuity, or Plus features.

Download and Install

Go to /download and grab the build for your platform.

Windows

1. Download the .exe installer.
2. Run it and follow the setup wizard.
3. Launch Sweet Tea Studio from the Start menu or your desktop shortcut.

If Windows shows a SmartScreen warning, click More info and then Run anyway.

Linux

You can use either build:

• AppImage if you want the fastest portable path.
• .deb if you prefer a package-managed install.

For AppImage:

chmod +x "Sweet Tea Studio-*.AppImage"
./"Sweet Tea Studio-*.AppImage"

For .deb:

sudo dpkg -i sweet-tea-studio_*.deb

Note: If you want the full reference chapter for install paths, data directories, or environment variables, jump to Installation & Setup.

Connect to ComfyUI

Sweet Tea needs a ComfyUI engine to actually run generation. You have two valid paths:

Path A: Let Sweet Tea manage it

This is the easiest path if you are starting fresh.

1. Open Settings.
2. Choose the managed ComfyUI option.
3. Point Sweet Tea at the install location you want to use.
4. Let Sweet Tea handle startup and connection for you.

Path B: Use the ComfyUI setup you already have

This is the right path if you already trust your own install.

1. Open Settings.
2. Enter the URL for your running ComfyUI instance, usually http://127.0.0.1:8188.
3. Confirm that Sweet Tea reports a healthy connection before you try to run anything.

Warning: Do not import a pile of workflows until you have one clean end-to-end generation working. Get the baseline stable first.

Make Your First Image

Once the engine is connected:

1. Open the main creation workspace.
2. Choose a Pipe. Start with one simple job, not the most complicated workflow you can find.
3. Type a straightforward prompt such as:

a quiet cabin in autumn fog, warm light in the windows, cinematic still

4. Leave most controls at their defaults.
5. Click Generate.
6. Wait for the run to finish.

Your first goal is not perfection. It is proof that the path works end to end.

Inspect the Result and Make One Variation

When the image finishes:

1. Open it from the run list or Gallery.
2. Check that the prompt, seed, model, and settings are attached the way you expect.
3. Make one small change and run again.

Good first variations are small:

• change one phrase in the prompt
• nudge the guidance or steps
• swap an input image
• reseed for a different composition

This is where Sweet Tea starts to matter. The result and the context that made it stay together.

Try a First Video or Motion Pass

Once your first image works, you can branch into motion work without changing the basic workflow:

1. Pick a Pipe meant for video, frame interpolation, or cleanup.
2. Reuse the same project so the new outputs stay attached to the same creative thread.
3. Start small: one short clip or one cleanup pass is enough for the first success.

If you do not need video right away, skip this and keep building confidence with still-image work first.

What to Do Next

• Understand the mental model — Core Concepts
• Tighten your install and settings — Installation & Setup
• Learn the full creation workspace — Prompt Studio
• Import your own workflows — Pipes & Workflows
• Need more VRAM later? — Hosted GPU Workflows

Next: Core Concepts