Agents

MCP server

Let an AI agent make videos for you. The Lanne Video MCP server works with Claude, ChatGPT, Claude Code, Cursor and any client that speaks Streamable HTTP. It uses your Studio credits: no API plan needed.

Server URL

MCP endpoint
https://lannevideo.com/mcp

Open it in a browser for a visual setup guide; agents talk JSON-RPC to the same URL. Your personal setup, with tokens and one-click buttons, is in Dashboard → MCP.

Authentication

Connect a client

Claude (desktop and claude.ai)

Customize → Connectors → Add custom connector, name it Lanne Video and paste the URL. Claude opens the consent screen.

ChatGPT

Settings → Connectors. Turn on Developer mode if asked, add a custom connector with the URL and approve access.

Claude Code and Cursor

claude mcp add --transport http lanne-video https://lannevideo.com/mcp \
  --header "Authorization: Bearer nbmcp_YOUR_TOKEN"

Then ask: "Make a 6 second vertical video of a barista pouring latte art."

Tools

ToolWhat it does
get_accountPlan, credits left and rate limit status.
list_modelsminimax-h3, seedance-2.5, flux-3 with their limits.
generate_text_to_videoPrompt in, video out. Arguments: prompt, model, duration, aspect_ratio, resolution. Waits up to 40 s, then returns video_url, or status: "processing" and a video_id for longer renders.
generate_image_to_videoPublic HTTPS image URL plus a motion prompt in, video out. Same wait and follow-up.
get_video_statusStatus and URL of a video_id. Pass wait_seconds (up to 50) to wait for a render to finish.
list_videosRecent generations on the account.
motion_*Edit Motion Studio projects live. The project has to be open in a browser tab; motion_list_projects, motion_get_project and motion_get_catalog work without it.

The server also ships prompt templates: ugc_product, cinematic_shot and animate_image.

Credits and limits

Agent generations use your Studio credits and the same rate limiter as the dashboard. When credits run out or the plan is paused, the tool returns an error the agent can relay to you.

Building your own product rather than using an assistant? The REST API gives you plain HTTP endpoints with the same models.