> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryflowy.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Flowy is a node-based AI creative platform: you generate images, video, audio, 3D and vector on an infinite Canvas, refine on the Studio timeline, and export or publish from the same project.
> Prefer the Flowy MCP server (https://mcp.tryflowy.ai/mcp) or the REST API at https://apis.tryflowy.ai/v1 for programmatic work. Install with `flowy mcp install` from the @flowy/cli package.
> Credits are workspace-scoped. Generations reserve credits on start and only deduct on success, so failed runs refund automatically.

# Elements

> Reusable characters, products, and locations that keep a subject consistent across generations.

**Elements** (`/library/characters`) are reusable reference subjects: a character, product, vehicle, logo, location, or anything else you want to look the same every time it appears. Build one once, then reuse it across as many projects as you like instead of re-describing it in every prompt.

<Frame caption="Elements: reusable reference subjects">
  <img src="https://mintcdn.com/flamapp/d-IQVAHPd1A_gxXA/images/library/elements.png?fit=max&auto=format&n=d-IQVAHPd1A_gxXA&q=85&s=5b5fdd3c4d3fa40848d307b5bc4724b2" alt="Elements" width="2880" height="1800" data-path="images/library/elements.png" />
</Frame>

<Note>
  Elements used to be called **Characters**: the page's URL and some older references still say `characters`, but the surface, the canvas node, and the API resource are all named Element.
</Note>

## Categories

Every element belongs to a category: Character, Creature, Object, Product, Vehicle, Logo, Environment, Location, or Other, or **Auto**, which lets Flowy infer the category from your reference images instead of picking one yourself. The library page's tabs filter by these same categories.

## Create an element

<Steps>
  <Step title="Choose a category">
    Pick the category that fits, or leave it on Auto and let the compiler infer it from what you upload.
  </Step>

  <Step title="Add reference images">
    Upload at least 1 and up to 4 images of the subject. More angles give the compiler more to work with.
  </Step>

  <Step title="Name it and add optional details">
    Give it a name, optionally steer how the compiler should treat the references (for example, ignoring an accessory), and optionally record a voice sample if the element should also speak consistently.
  </Step>
</Steps>

## Compiling the reference sheet

Once created, Flowy's reference-sheet compiler analyzes your images and builds a production sheet: multiple consistent views of the subject rather than the source photos alone. An element moves through a short lifecycle while this happens: **draft** while it's waiting to compile, **parsing** and **rendering** while the compiler works, then **review** once a sheet is ready for you to check. Approve the sheet to mark it **ready**, or recompile it if the result needs another pass, recompiling re-analyzes the references and rebuilds the sheet from scratch, so you'll render and approve it again afterward.

## Use an element in a generation

Add an **Element node** to your canvas, its keyboard shortcut is <kbd>⇧K</kbd>, and connect it into a workflow the same way you would any reference. Generations that use it draw on the compiled sheet with identity lock, keeping the subject's appearance consistent across outputs. How much of the sheet a generation can draw on depends on the output type: image generations use up to 4 sheet views, video and 3D/Flam generations use up to 3.

<Tip>
  Elements are also available over the API as `/v1/elements`, if you're integrating rather than working from the canvas.
</Tip>