MCP Server

Vidocu provides a Model Context Protocol (MCP) server that lets AI assistants like Claude interact with the Vidocu API through natural conversation.

Available on every plan

The MCP integration works on any Vidocu plan when you connect with OAuth - it's not limited to Business like API keys are. Operations draw from your plan's monthly video-minute allowance. (Connecting with an API key instead still requires the Business plan, since API keys do.)

Authentication

The MCP server supports two authentication methods:

Option 1: OAuth (recommended)

MCP clients that support OAuth (like Claude Desktop) can authenticate automatically. The user authorizes Vidocu through a browser-based consent flow — no manual key management needed.

The MCP server advertises Vidocu's OAuth authorization server via RFC 9728 Protected Resource Metadata at /.well-known/oauth-protected-resource.

{
  "mcpServers": {
    "vidocu": {
      "url": "https://mcp.vidocu.ai/"
    }
  }
}

The MCP client handles the OAuth flow automatically — redirects the user to Vidocu's consent screen, obtains tokens, and sends them as Authorization: Bearer on each request.

Option 2: API Key

For MCP clients that don't support OAuth, or for quick testing, pass your API key as a Bearer token:

{
  "mcpServers": {
    "vidocu": {
      "url": "https://mcp.vidocu.ai/",
      "headers": {
        "Authorization": "Bearer vdo_live_your_key_here"
      }
    }
  }
}

Get your API key from the Vidocu Dashboard.

Server details

URL	https://mcp.vidocu.ai/
Transport	Streamable HTTP
Authentication	Authorization: Bearer (API key or OAuth token)
OAuth metadata	https://mcp.vidocu.ai/.well-known/oauth-protected-resource

Available tools

The MCP server exposes 21 tools covering all Vidocu API operations:

Video management

Tool	Description
get_upload_url	Get an upload URL for a local video file
upload_video	Upload a video by providing a public URL
list_videos	List videos with optional filtering and pagination
get_video	Get details of a specific video

Project folders

Tool	Description
create_project	Create a project folder to organize videos
list_projects	List project folders with search and pagination
get_project	Get details of a specific project folder
update_project	Update a project folder's name or metadata
delete_project	Delete a project folder (optionally its videos too)

Operations

Tool	Description
analyze_video	Start video analysis to generate subtitles and metadata
export_video	Export a video with burned-in subtitles
translate_video	Translate subtitles to another language
generate_voiceover	Generate AI voiceover from subtitles
generate_article	Generate a help article from the transcript
get_subtitles	Get subtitles in JSON or SRT format
process_video	One-shot pipeline: analyze, voiceover, article, export, and translate in one step

Tools & utilities

Tool	Description
list_tools	Discover available video processing tools
get_tool	Get a tool's parameter schema
execute_tool	Run a video processing tool

Status & usage

Tool	Description
get_job_status	Check progress of an async operation
get_usage	View current plan, usage, and limits

OAuth scopes

When using OAuth, the MCP server requests the following scopes based on which tools are used:

Scope	Tools
videos:read	list_videos, get_video
videos:write	get_upload_url, upload_video
projects:read	list_projects, get_project
projects:write	create_project, update_project, delete_project
subtitles:read	get_subtitles
analyze:write	analyze_video
export:write	export_video
translate:write	translate_video
voiceover:write	generate_voiceover
article:write	generate_article
tools:read	list_tools, get_tool
tools:write	execute_tool
jobs:read	get_job_status
usage:read	get_usage
process:write	process_video

See OAuth Scopes for full details.

Async operations

Some operations (analyze, export, translate, voiceover, process, execute tool) run asynchronously. These tools return a jobId immediately. Use get_job_status to check progress:

You: "Analyze this video: https://example.com/demo.mp4"

Claude: I'll upload and analyze the video.
→ calls upload_video
→ calls analyze_video → returns { jobId: "analysis_..." }

Claude: The analysis is running. Let me check the status.
→ calls get_job_status → returns { status: "completed", result: { ... } }

Claude: Analysis complete! The video has 24 subtitles generated.

Example conversations

Generate subtitles:

"Upload this video and generate English subtitles: https://example.com/demo.mp4"

Export with subtitles:

"Export my latest video with subtitles burned in"

Translate:

"Translate the subtitles for video vid_abc123 to Spanish and French"

Full pipeline:

"Process this video end-to-end — analyze it, generate a help article, and export: https://example.com/tutorial.mp4"