AdSkull MCP server
AdSkull MCP · v0.1

Plug AdSkull into your AI.
Generate. Not just chat about it.

The official AdSkull Model Context Protocol server. Connect Claude, Cursor, Windsurf, Continue, or any MCP-compatible agent — and let it generate real images, videos, voiceovers, and talking avatars on your account, with your API key.

POSThttps://mcp.adskull.io/mcp

Auth: Authorization: Bearer YOUR_ADSKULL_API_KEY · Get a key at app.adskull.io/api.

It actually generates

Not a docs proxy. The model uploads assets, kicks off jobs, polls until done, and hands back finished URLs.

Every product, one server

Images, videos (text + image-to-video + extend), TTS, voice cloning, and lipsync talking avatars — 15 tools.

Your key, your credits

Multi-tenant by design. Every request is scoped to the Bearer key the agent sends — no shared secret.

Setup

Connect it to your favorite agent.

Drop this into your MCP client config. Replace YOUR_API_KEY with a key from your AdSkull dashboard.

// ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)
// %APPDATA%\Claude\claude_desktop_config.json (Windows)
{
  "mcpServers": {
    "AdSkull": {
      "type": "http",
      "url": "https://mcp.adskull.io/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}
// Settings → Features → MCP Servers → Add new
{
  "mcpServers": {
    "AdSkull": {
      "url": "https://mcp.adskull.io/mcp",
      "headers": { "Authorization": "Bearer YOUR_API_KEY" }
    }
  }
}
// ~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "AdSkull": {
      "serverUrl": "https://mcp.adskull.io/mcp",
      "headers": { "Authorization": "Bearer YOUR_API_KEY" }
    }
  }
}
// ~/.continue/config.json
{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "http",
          "url": "https://mcp.adskull.io/mcp",
          "headers": { "Authorization": "Bearer YOUR_API_KEY" }
        }
      }
    ]
  }
}
# Streamable HTTP MCP transport
URL:    https://mcp.adskull.io/mcp
Method: POST (JSON-RPC 2.0)
Auth:   Authorization: Bearer YOUR_API_KEY

Tools

15 tools your model can call.

Each tool is documented inline so the LLM knows when and how to use it. Asset uploads happen automatically when a local file or external URL is provided.

upload_asset

Host an image, audio, or video file so other tools can reference it.

generate_image

Text-to-image with optional reference images.

generate_video

Text or image-to-video. Polls to completion.

extend_video

Lengthen a previously generated clip.

generate_voice

TTS or speech-to-speech with any voice.

generate_talking_avatar

Lipsync a portrait to a script or audio file.

clone_voice · design_voice

Build a custom voice from audio or from a description.

list_custom_voices · delete_custom_voice

Manage your voice library.

list_models / list_voices / list_pricing

Discover what's available and what it costs.

get_credits / get_usage / get_plan

Check quota and plan before expensive jobs.

get_job_status

Re-check a job by ID if a generation timed out.

Try it

Ask your agent something like this.

"Generate a 1:1 hyperreal product photo of a glowing energy drink,
then animate it into a 5-second slow rotation video with studio
lighting, and save the final video URL."

Your model will pick generate_image, then upload_asset the result, then generate_video with that hosted URL — all on its own. You just get the final link.