OpenCode Integration with Luma MCP

📘 Overview：OpenCode MCP Overview →

Luma Dream Machine has always been one of the most cost-effective video models in terms of “low cost / high image quality / natural transitions.” After connecting the Luma MCP Server in OpenCode, you can have it generate short videos based on text or reference images, and you can also directly extend existing videos while maintaining stylistic consistency.

Luma MCP is suitable for product demo short clips, Banner animations, social media short videos, as well as two high-frequency scenarios: “driving animation with images” and “video extension.”

Get API Token

Before using the Luma MCP Server, you need to prepare an Ace Data Cloud API Token. OpenCode and Claude Code share the same Token, and the acquisition process is the same:

  1. Open Ace Data Cloud Console - Applications to obtain your API Token and keep it for later use.
  2. If you are not logged in or registered yet, you will automatically be redirected to the login page; after logging in or registering, you will automatically return to the current page.
  3. A free quota will be provided when you apply for the first time, so you can try the Luma MCP service for free first.

Get Ace Data Cloud API Key

One Token can be used with all MCP Servers provided by Ace Data Cloud, so there is no need to apply separately for Luma. In documents and screenshots, it is recommended to display only a redacted form, such as YOUR_ACEDATACLOUD_API_KEY, and do not paste the complete Token into public repositories, Issues, screenshots, or chat records.

Want to use Claude Desktop / Claude.ai web version for one-click OAuth authorization directly? Please see the Luma MCP Claude.ai / Desktop Tutorial.

Configure OpenCode

OpenCode uses an opencode.json file to describe all MCP Servers. The file can be placed in two locations; choose one according to the “scope” below. The field structure of the two locations is exactly the same, with the only difference being their priority.

It is strongly recommended to write the Token into an environment variable first:

export ACEDATACLOUD_API_KEY="Replace with your actual Token"

Then use {env:ACEDATACLOUD_API_KEY} as a placeholder reference in opencode.json to avoid writing the actual Token into the file.

Global: Shared by All Projects

File location: ~/.config/opencode/opencode.json, suitable for “my own machine, configure it once, and use it for all projects.”

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "luma": {
      "type": "remote",
      "url": "https://luma.mcp.acedata.cloud/mcp",
      "enabled": true,
      "oauth": false,
      "headers": { "Authorization": "Bearer {env:ACEDATACLOUD_API_KEY}" }
    }
  }
}

⚠️ "oauth": false cannot be omitted. AceData’s MCP Server uses Bearer Token authentication and does not use the OAuth flow; OpenCode will automatically treat 401 as an OAuth challenge and redirect by default, causing opencode mcp list to show SSE error: Non-200 status code (401). Explicitly declaring "oauth": false allows OpenCode to call directly with Authorization: Bearer ...; opencode mcp debug luma will also echo OAuth explicitly disabled, indicating that it has taken effect.

Project-Level: Effective Only for the Current Project

File location: opencode.json under the root directory of the current project, which will override the global configuration. Suitable for team sharing, single-repository customization, or temporarily enabling an MCP only in a certain project.

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "luma": {
      "type": "remote",
      "url": "https://luma.mcp.acedata.cloud/mcp",
      "enabled": true,
      "oauth": false,
      "headers": { "Authorization": "Bearer {env:ACEDATACLOUD_API_KEY}" }
    }
  }
}

If the project-level opencode.json will be committed to Git, make sure to use the {env:ACEDATACLOUD_API_KEY} placeholder instead of the actual Token to avoid leakage.

💡 If .env in the Shell does not have an explicit export, you need to use set -a && source .env && set +a to pass the variable to the OpenCode process; otherwise, opencode debug config will show "Authorization": "Bearer " (the placeholder was not resolved).

Verify Connection

After configuration, run:

opencode mcp list
opencode mcp debug luma

Seeing luma connected means the handshake was successful. If it fails, check whether ACEDATACLOUD_API_KEY has been exported, whether oauth is false, and whether Authorization contains the Bearer prefix. Model and tool Schemas will continue to change, so select a model that supports native tool calling from the current directory.

Typical Scenarios

After configuration is complete, you can directly use natural language in the OpenCode session to make calls without first switching to /mcp:

Text-to-Video

Generate a 5-second video: waves gently lapping against a golden beach, sunset, cinematic feel

Image-to-Video

Turn this product image into a video, slowly circling around it from left to right

Video Extension

Extend this video, add another 10 seconds, and maintain the same style

Tool List

The table below lists the main tools; you can obtain the current complete list through curl -X POST https://luma.mcp.acedata.cloud/mcp -H 'Authorization: Bearer <token>' -H 'Accept: application/json' -H 'Content-Type: application/json' --data '{"jsonrpc":"2.0","id":1,"method":"tools/list"}':

Tool Description
luma_generate_video Text-to-video
luma_generate_video_from_image Generate video using a reference image
luma_extend_video Video extension