Skip to main content
MCP overview

MCP overview

Create content assets from any MCP-compatible AI assistant.

What is MCP?

MCP (Model Context Protocol) lets AI assistants like Claude, Cursor, Windsurf, and Codex call external tools. AgentFramer exposes content creation tools through MCP, so your assistant can generate images, video, and audio without custom API work.

How it works

Connect AgentFramer to your MCP client once. After that, your assistant can pick models, start generation jobs, check status, fetch output URLs, and continue working with recent assets.

MCP and the dashboard use the same workspace. Assets generated by an agent are stored in AgentFramer, visible in the dashboard, and billed against prepaid workspace balance.

What you can ask your AI to do

With MCP connected, your assistant can create and manage content assets in plain language. Examples:

Setting it up

  1. Run npx agentframer@latest init or install AgentFramer from your MCP client.
  2. Sign in and authorize the connection when AgentFramer opens the browser.
  3. Add prepaid balance during onboarding or from workspace billing.
  4. Restart the MCP client if required. The AgentFramer tools appear in your assistant.

Workspaces and billing

Your MCP session is scoped to one workspace. Generation history, team access, and prepaid balance are all workspace-scoped, so agents can work inside the same project context as the dashboard.