MCP · 11 platforms

Your AI can now read video

Name: Video Intelligence MCP
Author: Video Intelligence MCP

Video Intelligence MCP — one remote MCP endpoint for transcripts, raw subtitles, metadata, chapters, and playlist jobs from public video platforms.

Join early access Share your use case View on GitHub

structuredContent

1{
2  "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
3  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
4  "uploader": "Rick Astley",
5  "duration_sec": 213,
6  "tools": [
7    "get_video_info",
8    "get_transcript"
9  ],
10  "transcript_preview": "♪ We're no strangers to love ♪ ♪ You know the rules and so do I ♪ ♪ A full commitment's what I'm thinking of ♪ ♪ You wou…"
11}

Works with

Claude Desktop

Open Claude Desktop → Settings → Developer → Edit Config. Add transcriptor as a custom connector with your remote MCP URL. Local stdio (Docker) — coming soon.

Open Settings → Developer → Edit Config
Add the transcriptor server block with your remote MCP URL
Restart Claude Desktop and verify tools appear in chat

Config

mcp config

1{
2  "mcpServers": {
3    "transcriptor": {
4      "url": "https://your-mcp-endpoint.example.com/mcp"
5    }
6  }
7}

Cursor

Open Cursor Settings → MCP. Add transcriptor with a remote URL in global config or project `.cursor/mcp.json`. Local stdio (Docker) — coming soon.

Open Cursor Settings → MCP
Edit global mcp.json or create `.cursor/mcp.json` in your project
Set the transcriptor server URL to your remote MCP endpoint
Reload the window — tools show up in Agent and Composer

Config

mcp config

1{
2  "mcpServers": {
3    "transcriptor": {
4      "url": "https://your-mcp-endpoint.example.com/mcp"
5    }
6  }
7}

Claude Code

Add transcriptor as a remote MCP server from your project root (`claude mcp add` with HTTP URL), or put the URL in `.claude/settings.json`. Local stdio (Docker) — coming soon.

Run `claude mcp add` with your remote transcriptor MCP URL
Confirm the server in `claude mcp list`
Invoke tools from Claude Code sessions in your repo

Config

mcp config

1{
2  "mcpServers": {
3    "transcriptor": {
4      "url": "https://your-mcp-endpoint.example.com/mcp"
5    }
6  }
7}

OpenClaw

Add transcriptor under mcp.servers in ~/.openclaw/openclaw.json with a remote streamable-http URL. Local stdio (Docker) — coming soon.

Edit ~/.openclaw/openclaw.json and add the transcriptor remote server URL
Restart the gateway if tools do not appear (openclaw gateway restart)
Verify with openclaw mcp doctor — get_transcript should be listed

Config

mcp config

1{
2  "mcpServers": {
3    "transcriptor": {
4      "url": "https://your-mcp-endpoint.example.com/mcp"
5    }
6  }
7}

n8n

In n8n, add an MCP Client Tool node or connect transcriptor to an AI Agent node via your remote MCP endpoint URL.

Add an MCP Client Tool or AI Agent node to your workflow
Set the MCP server URL to your transcriptor endpoint
Wire tool outputs into downstream automation steps

Config

mcp config

1https://your-mcp-endpoint.example.com/mcp

Coverage

11 platforms · 7 MCP tools

Synced with transcriptor-mcp — URL-based tools work on all platforms; search_videos is YouTube only.

Platforms

Tutorial → study notes

Paste a course URL; your agent summarizes chapters and pulls key quotes with approximate timestamps.

get_video_chapters
get_transcript
get_video_info

MCP tools

Transcript without timestamps or HTML — ready context for LLMs.
Example
usage
1Call get_transcript for https://youtube.com/watch?v=…, lang "en". 2Return a 5-bullet summary of the main points.
Source subtitles with timestamps; safe for long videos.
Example
usage
1{ "url": "https://youtube.com/watch?v=…", "format": "srt", "response_limit": 1000 }
Discover available tracks before choosing type/lang.
Example
usage
1get_available_subtitles for the URL → list tracks. 2Pick official "en" if available, else auto-generated.
Title, channel, tags, thumbnails, viewCount, and more via yt-dlp JSON.
Example
usage
1{ "url": "https://youtube.com/watch?v=…" } 2→ title, uploader, duration_sec, tags

Video structure for navigation and summarization.

Example

usage

1get_video_chapters + get_transcript for the URL.
2Convert chapters into a blog outline with H2 per chapter.

Process playlist URLs in a single tool call.
Example
usage
1get_playlist_transcripts for a playlist URL. 2Output a table: video title | theme | 3 bullet takeaways.
Search YouTube — the only platform-specific tool.
YouTube only
Example
usage
1{ "query": "react hooks tutorial", "limit": 5 } 2→ pick the best match, then get_transcript

Time to value

Three steps to video context

From MCP connection to structuredContent — no REST glue code.

01
Connect MCP
Claude custom connector, Cursor, or self-hosted stdio + mcp-proxy. Public remote MCP endpoint — on the roadmap.
02
Paste a URL
Any public URL from 11 supported platforms — or a YouTube video ID.
03
Receive structuredContent
Transcript, raw subtitles, metadata, chapters — typed JSON for your agent.

Capabilities

Transcript, metadata, chapters — via MCP

7 typed tools with structuredContent for agent workflows.

Transcript & raw subtitles
Cleaned text or SRT/VTT; multi-language; Whisper fallback when subtitles are missing.
11 platforms
YouTube, X, Instagram, TikTok, Twitch, Vimeo, Facebook, Bilibili, VK, Dailymotion, Reddit.
Metadata & chapters
Title, channel, tags, thumbnails, duration, chapter markers — structuredContent for agents.
Playlist batch
get_playlist_transcripts for bulk context extraction from playlists.
Pagination
Safe pagination for large transcripts via response_limit and next_cursor.
Text and metadata only
Text and metadata only — no media file downloads.

Use cases

For developers, agencies, content, and analysts

One MCP endpoint — different prompt-driven workflows.

MCP-native workflow in minutes

7 typed tools with structuredContent — no REST glue code
Docker-first: stdio locally or mcp-proxy for HTTP/SSE
Pagination for large transcripts out of the box
Prometheus metrics on REST API (self-hosted)

Developers prompt

1"Use Video Intelligence MCP.
2First call get_video_info with the URL.
3Then call get_transcript in English.
4Return structured summary plus timestamps."

Research and competitive intel from video

Batch playlist transcripts for client deliverables
Extract quotes with timestamps for reports
Multi-platform — not just YouTube
Team workspaces — planned post-launch

Agencies prompt

1"Analyze this YouTube playlist.
2Return:
31) 10 main themes
42) recurring claims
53) quotes with timestamps
64) a table: video → theme → key takeaways"

Repurpose video into text assets

Transcript → blog posts, newsletters, social clips
Chapters for outlines and timestamps
Raw subtitles for precise quoting
Multi-language subtitle discovery

Content prompt

1"Convert this video transcript into:
2- 5 tweet hooks with timestamps
3- A blog outline with H2 sections
4- Key quotes for LinkedIn"

Due diligence and fact extraction

Structured metadata: uploader, views, upload date
Entity and claim extraction via agent prompts
Playlist-level pattern analysis
Public content only — no private media access

Analysts prompt

1"Extract all named entities, financial claims, and forward-looking statements from this earnings call video. Include timestamps for each claim."

Pricing

Free to start, Indie when you ship

Launch tiers for solo devs — join early access while billing is on the roadmap.

Free

$0/mo

Try Video Intelligence MCP before you commit.

50 credits/month
Single video transcripts
Basic metadata
Low rate limit
No batch playlist transcripts

Indie

$9/mo

For solo developers shipping with Claude, Cursor, or ChatGPT.

500 credits/month
All core tools
ChatGPT / Claude / Cursor setup
Basic playlist support

Launch tiers above are early-access guidance. Final limits may adjust before public billing. Stripe checkout is on the roadmap.

View full pricing Join early access

FAQ

Frequently asked questions

Compatibility, credits, public content, auth, and the deep research caveat.

Yes.

Yes, with caveats: public videos with available subtitles or Whisper fallback only. No access to private content, DRM, or paywalled media. Long playlists use pagination and batch tools.

No. Subtitle text and metadata via yt-dlp only. Media files are not saved to disk for the user.

Ask us for Whisper fallback support.

Join early access

Waitlist lead qualification. Self-hosted Docker is on GitHub today; public remote MCP endpoint — on the roadmap.

Your AI can now read video

Claude Desktop

Platforms

Tutorial → study notes

MCP tools

Connect MCP

Paste a URL

Receive structuredContent

Transcript & raw subtitles

11 platforms

Metadata & chapters

Playlist batch

Pagination

Text and metadata only

MCP-native workflow in minutes

Research and competitive intel from video

Repurpose video into text assets

Due diligence and fact extraction

Free

Indie