Panda Video can connect to AI tools like Claude Code, Claude Desktop, Cursor, and Windsurf through the Model Context Protocol (MCP), an open source standard for AI-tool integrations. The Panda Video MCP server exposes 128 tools covering the full Panda Video API — videos, folders, playlists, profiles, lives, analytics, funnels, watermark, AI features, and more.

Connect the MCP server when you find yourself copying video IDs, folder names, or analytics data into your AI assistant. Once connected, the AI can read and act on your Panda Video account directly instead of working from what you paste.

What you can do with the Panda Video MCP

With the MCP server connected, you can ask your AI assistant to:

Manage videos: "List my videos uploaded in the last 7 days and rename the ones with placeholder titles."
Organize content: "Create a folder called 'Q1 Tutorials' and move videos 12345 and 67890 into it."
Analyze performance: "Show me the retention curve of video X and compare it with video Y."
Manage live streams: "Create a new live stream key called 'Webinar March' and return the RTMP URL."
Automate AI features: "Generate AI subtitles in Portuguese and Spanish for all videos in the 'Onboarding' folder."
Run analytics queries: "What were the top 10 videos by unique plays last month, broken down by country?"
Manage funnels: "Add this video to the 'Sales Funnel 2026' with coordinates at 30 seconds."

Requirements

Node.js 18 or higher
A Panda Video API Key — get it in Settings → Advanced → API Key

🚧
Treat your API key like a password
The MCP server runs with full account privileges. Only connect it to AI clients you trust, and never commit your API key to source control.

Quick start

Get up and running in under 2 minutes using Claude Code.

Get your API key

In the Panda Video dashboard, go to Settings → Advanced → API Key and copy the key. Keep it private — it has full account access.

Connecting to MCP clients

MCP servers can run as a local stdio process or as a remote HTTP server. Use stdio when the client runs on the same machine as the server; use HTTP when you want one server to serve multiple clients or a remote IDE.

Claude Code

Run this in your terminal (replace YOUR_KEY with your actual key):

claude mcp add panda-video --transport http https://mcp.pandavideo.com/mcp --header "Authorization: panda-your-key"

▎ The panda- prefix is literal — paste your key right after it, with no space.

Open Claude Code and ask:

▎ "Using the panda-video MCP, list my 5 most recent videos."

If Claude returns video titles from your account, you're connected. If you get a 401 Unauthorized, re-check the key.

Claude Desktop (stdio)

Edit claude_desktop_config.json:

{
  "mcpServers": {
    "panda-video": {
      "command": "npx",
      "args": ["-y", "panda-video-mcp"],
      "env": { "PANDA_API_KEY": "your-key" }
    }
  }
}

Cursor / Windsurf (HTTP)

Add to your MCP settings:

 {
    "mcpServers": {
      "panda-video": {
        "type": "http",
        "url": "https://mcp.pandavideo.com/mcp",
        "headers": {
          "Authorization": "panda-your-key"
        }
      }
    }
  }

Replace your-key with your Panda API key.

Claude Desktop / Gemini CLI (HTTP)

 {
    "mcpServers": {
      "panda-video": {
        "type": "http",
        "url": "https://mcp.pandavideo.com/mcp",
        "headers": {
          "Authorization": "panda-your-key"
        }
      }
    }
  }

Replace your-key with your Panda API key.

Tools (128 total)

The server groups tools by domain so the AI client can pick the right one without flooding the context window.

Domain	Count	Tools
Videos	7	list, get, update, recover, delete, upload URL, upload m3u8
Folders	10	CRUD, library info, qualities, referer domains
Playlists	8	CRUD, video management, password-protected access
Profiles	6	CRUD, apply to videos (colors, autoplay, smart autoplay)
Live	13	CRUD, finish, metrics, stream keys, chat CRUD
Analytics	14	account, video, retention, ranking, geo, browser, referrer, realtime, live, viewers, bandwidth, billing, users, credits
Funnels	7	CRUD, add/remove videos with coordinates
Review	11	Review links, timestamped comments (owner + public APIs)
Watermark / DRM	14	Status, groups CRUD, JWT, private tokens CRUD
Thumbnails	2	get, delete
Subtitles	4	get, get by language, delete, AI generate
Custom Fields	4	CRUD
Organization	2	get, update
Users	5	CRUD
Downloads	3	request async, check status, direct download
Shorts	6	list jobs, list segments, delete, save to library, download
AI Assistants	8	CRUD, video bindings, chat visibility, purchase overview
AI Dubbing	1	create dubbing job
Referrals	3	summary, transactions, send invites

MCP resources

The server exposes two MCP resources you can reference with @ in your AI client:

panda://api/limits — Rate limits, batch limits, plan constraints
panda://api/overview — Tool listing by domain

Limits and security

Rate limiting: 60 requests/minute per API key (token bucket)
Batch limits: Max 50 items per batch operation
API key validation: Validated at startup (stdio) or on first tool call (HTTP)
Error handling: All API errors mapped to LLM-friendly messages
Pagination: All list endpoints support page and per_page (max 100)

Practical examples

Manage videos across folders

List all videos in the "Drafts" folder that haven't been updated in 30 days, then move them to the "Archive" folder.

Analyze retention before publishing

Get the retention curve for video 12345 and tell me at which timestamps viewers are dropping off most sharply.

Generate AI subtitles in bulk

For every video in the "Onboarding 2026" folder, generate AI subtitles in pt-BR, en, and es.