VideoNote-MCP
An MCP server that turns a video or podcast link into a transcript, subtitles and a Markdown note, with frames and speech.
- 103 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Downloads files by link and runs FFmpeg and speech recognition on your machine.
- Stores platform cookies and tokens in a local config.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/videonote-mcpDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Downloads files by link and runs FFmpeg and speech recognition on your machine.
- Stores platform cookies and tokens in a local config.
This is third-party code. Review the repository files before installing.
What it does
VideoNote-MCP gathers note material from a video or podcast link or from a local file. The server downloads the clip, takes the platform's official subtitles and falls back to speech recognition when they are missing. On request it extracts key frames, comments and Bilibili danmaku. The `prepare_note_material` tool hands the material to the current assistant, which writes the note without a separate LLM key. `generate_note` plugs in your own model as a fallback and `process_media` exports subtitles as SRT, VTT and JSON. Sources include Bilibili, YouTube, Douyin, TikTok, Kuaishou, Xiaohongshu and Xiaoyuzhou podcasts.
Who it is for. Students and writers who take notes on lectures, podcasts and video essays with an agent and want Markdown with a transcript and screenshots.
Good fit when
- A lecture or course on Bilibili or YouTube becomes a note with frames.
- A meeting recording is merged from parts, split by speaker and formatted as minutes.
- Clips need SRT or VTT subtitles for another tool.
- A Xiaoyuzhou podcast is summarized from the official transcript without a long local transcription.
Not a fit when
- The video sits behind a login and signing in to the platform is not possible: several services need cookies.
- You need a finished PDF: the server does not compile LaTeX or Typst, it supplies material and templates.
- The computer has no FFmpeg or uv and you cannot install them: the installer does not add them.
Example request
Take a Bilibili lecture link, prepare material with frames and write a Markdown note with chapters and a conclusions section.MCP
- Transport
- stdio
- Authentication
- not required
Environment variables
VIDEONOTE_DATA_DIR- Data directory for tasks, notes and frames.
VIDEONOTE_CONFIG_DIR- Directory for settings, logins and models.
Limitations
Python 3.11 to 3.13, uv and FFmpeg are required and installed beforehand. Some platforms (Bilibili AI subtitles, Douyin, Xiaohongshu) need a login: `videonote login <platform>` stores a cookie or token locally, so a separate account is wise. faster-whisper is installed by default, while mlx-whisper, FunASR and speaker diarization are added with `--with`, identical at install and in `setup`. Long recordings take a while to transcribe. Documentation and the install wizard are mostly in Chinese, and a README_EN exists.
How to disable. Remove the registration through the `videonote setup` wizard or your client's MCP removal command. Data lives in `~/.local/share/videonote-mcp/` and is cleaned by the `cleanup` tool or by hand.
FAQ
Is an LLM key required?
No. By default the connected assistant writes the note from the material returned by `prepare_note_material`. A key is only needed for the `generate_note` fallback.
What if a platform asks for a login?
Run `videonote login bilibili` or the command for the platform you need. QR-code sign-in or pasted cookies are stored locally only.
Where do results go?
To the `note_results/{task_id}/` folder: `raw/` with sources and `gen/` with the transcript, note and frames.
Which sources are supported?
Bilibili, YouTube, Douyin, TikTok, Kuaishou, Xiaohongshu, Xiaoyuzhou podcasts and local audio and video files.
Which export formats exist?
SRT, VTT and JSON subtitles and a Markdown note. The server supplies LaTeX and Typst templates but does not compile them.
Discussion
No questions or reviews yet.
Related
MCP serversMCP server, CLI and desktop app that cut a long video or stream replay into vertical clips with captions
VidMoat MCP
Vidmoat MCP
Remote MCP server of a cloud video editor: an agent imports footage, edits the timeline, checks frames and renders MP4
MCP server, /video plugin and CLI: transcript, key frames, on-screen text and metadata for a YouTube, Loom or local video
A skill that turns a Bilibili, Douyin or YouTube video link into timestamped notes through local transcription
ClipForge MCP
ClipForge
ClipForge MCP builds a vertical video from a topic or a product link: stock footage, Edge TTS voiceover and captions, no keys, composed with FFmpeg
MCP server that gives an agent a YouTube video's transcript from its link, with language choice, timestamps and ads filtered out
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| VideoNote-MCP (this card) | no | not required | 103 | |
| video-to-notes | no | 113 | builds notes from video with another pipeline; VideoNote-MCP is an MCP server with platform subtitles and frames | |
| yt-dlp-mcp | yes | not required | 309 | downloads video and subtitles; VideoNote-MCP adds recognition, frames and note material |
README in short
The README presents a pipeline from video link to note as an MCP server with no separate backend. It covers inputs (links, local files, ready subtitles), output formats, two worked examples and task management through `list_tasks`, `cleanup` and `health_check`.