# Layer API documentation

> Programmatic access to 500+ generative AI models for game-ready 2D, 3D, video, and audio assets, all from a single API.

## Generate any modality

One API for image, video, 3D, and audio: 500+ models plus the custom styles your team trains.

[![Image generation](/docs/_astro/image.Bx9xRZXt_ZUcPw6.webp)Image](/docs/models)[![Video generation](/docs/_astro/video.aU6q54_b_ZqPbQd.webp)Video](/docs/models)[![3D model generation](/docs/_astro/3d.CdLcapKK_2v1RJt.webp)3D](/docs/models)[![Audio generation](/docs/_astro/audio.CsFakVKB_Z1SKQyR.webp)Audio](/docs/models)

## Key concepts

**Workspaces** are the top-level container. Every API call is scoped to a workspace, which holds your team, models, assets, and Creative Unit balance.

**Models** are the AI engines that generate assets. Each supports specific modalities (image, video, 3D, audio) and capabilities (reference images, inpainting, style transfer), with 500+ built-in plus any custom styles your team has trained.

**Inferences** are individual generation runs. You pick a model, provide a prompt and parameters, and Layer returns the assets. Inferences run asynchronously: you start one and poll for results.

**Workflows** are multi-step pipelines built in the Layer app’s Blueprint editor. They chain operations (e.g. generate an image, upscale it, remove the background) and are callable as a single API endpoint.

**Files** are uploaded inputs (reference images, masks, poses, audio) that you pass to inferences as guidance.

**Creative Units (CUs)** are the billing currency. Cost varies by model and output size. Use the estimate endpoints to check cost before generating.

## Start building

[Quickstart](/docs/quickstart)Generate your first asset end-to-end.

[Authentication](/docs/authentication)Create a token and make your first authenticated call.

[CLI](/docs/cli)Generate from your terminal, or script Layer into a pipeline.

[SDK](/docs/sdk)Generate from JavaScript and TypeScript, with types and retries built in.

[MCP server](/docs/mcp)Use Layer directly from agent conversations.

[REST API](/docs/v2/rest-api)Full endpoint documentation with interactive examples.