# Getting Started

Framo is a workspace for generating, transforming, and composing AI images, videos, and 3D models on an infinite canvas, with a 2D pixel editor and a 3D scene editor built in.

## The big picture

A project opens in the **editor** (the canvas + top bar + a left icon rail). The left rail (only visible inside a project) has three surfaces you switch between:

- **Scene** (Box icon, or press `1`) — the 3D scene editor.
- **AI 1.0** (Sparkles icon, or press `2`) — the infinite AI canvas with the full sidebar.
- **2.0** (Sparkles "2.0" icon) — chat mode: the infinite canvas with the floating chat bar / agent instead of the sidebar.

The top bar shows your project name (centre), your credit balance / plan (right), and a Dashboard link (Framo logo, top-left).

## The infinite canvas

The infinite AI canvas is where your generations live as **nodes** you can move, select, and connect. Each image, video, or 3D result lands here.

To work on the canvas:
- **Select** a node by clicking it. What you select drives the AI's mode (see below).
- **Multi-select** by holding `Shift` and clicking, or by dragging a selection box around several nodes.
- **Add your own image:** right-click anywhere on the canvas to open the context menu, then **Upload Image** or paste from clipboard.
- **Canvas settings** live bottom-left (gear icon) — toggle info badges, node edges, video auto-play, auto-place, and trackpad scroll/zoom.
- An empty canvas shows a "Nothing here yet" card with a **Take a tour** button.

## Creating images, videos, and 3D models

You generate from the chat bar (2.0) or the AI sidebar (1.0). Pick a **mode**, pick an **AI Model**, write a **prompt**, then **Generate** (or `Ctrl+Enter`). The credit cost shows on the Generate button.

- **Image** — text-to-image with nothing selected; image-to-image with one image selected; multi-image with several selected.
- **Video** — animates images into clips. One selected image becomes the start frame; two images become start + end frames.
- **3D** — turns images into a 3D model. One image generates from a single view; up to 8 images combine multiple views for a more accurate model.
- **Enhance** — an AI render engine that takes your 3D scene or a live screen share as input and produces polished images.

Prompt tools sit on the right edge of the prompt input: Favorites, History, Presets, an AI prompt Enhancer, From Image, and Dictate (speak your prompt). See [Chat bar: generating images](/docs/product/chat-image-gen) for the full composer.

New images first arrive in the **Generation Collector** — swipe right to keep a result and add it to the canvas, or left to discard.

Everything you make is saved to the **Media Explorer** (right side). Search by prompt, filter by images/videos, and drag any item back onto the canvas to reuse it. See [Projects, media library & stock](/docs/product/projects-media).

> Steps chain together: a generated image can be selected as input for the next image, used as a video start frame, or turned into a 3D model. See the playbooks.

## The 2D (Pixi) editor

Double-click an image node on the canvas to open the **2D editor** — a Pixi-based pixel/vector editor for painting, masking, drawing paths, and editing layers on that image. The chat bar locks to image mode while you're inside. See [2D canvas editor](/docs/product/editor-2d).

> From here you can generate edits (e.g. inpainting on a drawn selection) without leaving the image.

## The 3D editor

The **Scene** tab (press `1`) is the Three.js 3D editor: build and arrange a 3D scene with objects and materials. Use the **Enhance** mode to render the live scene into finished AI images. See [3D scene editor](/docs/product/editor-3d).

> The Scene tab and the AI canvas share one project — switch between them on the left rail.

## Projects and saving

Your work lives in a **project**. The current project's name is in the centre of the top bar — double-click it to rename. The chevron beside it opens a menu to start a **New Project**, **Share Project**, or switch to another project. Saving is automatic (autosave); switching or creating a project saves first. All projects are listed in the Dashboard. See [Projects, media library & stock](/docs/product/projects-media).

## Getting help

- **Take a tour:** click the **?** (help) button bottom-right of the canvas, or the **Take a tour** button on an empty canvas — a ~1-minute guided walkthrough of the AI tools. You can resume it where you left off.
- **Ask the assistant:** press **⌘/Ctrl + K** and choose **Ask Framo Assistant**, then type your question (e.g. "how do I get from a sketch to a 3D model?"). It answers from this help knowledge base and is aware of what's on your current canvas, so it can suggest the concrete next step for your project.
- **Debug a failed generation with the assistant:** select the relevant image nodes first — for example your source image plus the failed attempts — then open the assistant (⌘/Ctrl + K) and ask what went wrong. It can *see* the selected images and read their prompts, tell you why an edit drifted or failed, and propose a better prompt you can apply with one click (it can also re-select the right nodes for you). Remove the attached images from the chip in the assistant header if you don't want them analysed.