The last mile of an asset pipeline is rarely a person clicking a button. It's a build script, a content management system, an internal tool — or increasingly, an AI agent — that needs assets on demand. If your generation platform only lives in a browser tab, every automated system upstream and downstream has to route around it.
Layer's public REST API and MCP server make the whole platform programmable. Everything your team does in the app — generating, training, running workflows, searching the library — is callable from your own code and your own agent stacks.
Key Capabilities
A REST API Across the Platform The API covers files, generations, models, workflows, training, projects, workspaces, members, and search. Automate generation, kick off training runs, query the library, and manage workspace structure from code.
Personal Access Tokens Authentication is straightforward — create and manage personal access tokens from your user settings, and scope your integrations to the people and systems that own them.
A Native MCP Server The Layer MCP server exposes Layer to external AI agents and tools. Agents can list and execute workflows, run generations, get model recommendations, start training, and upload files — Layer becomes a capability your agents call, not a tab they can't reach.
Workflows as Callable Services Pipelines your team builds and publishes on Layer's node canvas are executable through the API and MCP server. The art-approved process becomes an endpoint any internal system can invoke.
One Integration, 149+ Models A single API surface fronts Layer's models across image, video, 3D, and audio. When you adopt a new model, your integration doesn't change — the request does.
How It Works
Create a Token Generate a personal access token from your user settings and store it wherever your pipeline keeps secrets.
Call the API or Connect via MCP Hit the REST endpoints directly from scripts and services, or point your agent framework at the Layer MCP server to give agents the full toolset.
Automate the Pipeline Trigger generations from your CMS, execute published workflows from CI, feed results back into your build — Layer runs as a component of your system.
Built for Game Production
Content Pipelines That Run Themselves Live-ops calendars are automation problems. Drive event asset production from your own scheduling systems by executing published workflows on demand.
Agent-Ready by Design As studios build internal AI agents, the MCP server means Layer is already a tool those agents can use — generation, training, and workflow execution included — without custom glue code.
UA Creative on Tap Marketing systems can request variants programmatically and pull results straight from the API, keeping the creative-testing loop running without manual exports.
Enterprise Controls Underneath Programmatic access sits on the same foundation as the app — SOC 2 Type II, SSO/SCIM, role-based access, and exportable audit logs — and customer data is never used to train foundation models.
The teams getting the most out of AI treat it as infrastructure, not an app. The API and MCP server make Layer a building block in whatever you're building next.