Video to Notes
Video to Notes is a core new feature of Open Note: paste a video link or select a local audio/video file, and a structured note is generated automatically.
Highlights
- Multiple input methods: Paste a video link, or select local MP4 / MP3 / WAV audio/video files
- Offline local recognition: Speech recognition runs locally and offline — audio is never uploaded, keeping your data private and secure
- Structured output: Content is automatically organized into a structured note with a summary, key points, and to-do items
How to Use
Option 1: Paste a Video Link
- Copy the video link from a video platform
- Use the "Video to Notes" feature in Open Note and paste the link
- Wait for processing to finish — a structured note is generated automatically
Option 2: Import a Local Audio/Video File
- Select a local file in the "Video to Notes" feature; MP4 / MP3 / WAV and other common formats are supported
- Wait for local speech recognition to finish
- A structured note is generated automatically
Generated Content
Once processing is complete, Open Note automatically organizes the audio/video content into a structured note:
- Summary: Quickly understand the overall content
- Key points: Core knowledge and important information
- To-do items: Action items extracted from the content
The generated note supports further editing and can be exported as TXT / MD / PDF via the Import & Export feature.
Privacy
Speech recognition runs entirely locally and offline. Audio files are never uploaded to any server, so you can use it with confidence.
Use Cases
- 📹 Meeting recordings: Automatically generate meeting minutes and to-do lists
- 🎓 Online courses: Turn lectures into notes with clear key points
- 🎙️ Podcasts: Convert great audio content into text for easy review and search
- 📺 Talks & presentations: Quickly turn tech talks and speeches into lasting knowledge
FAQ
Installation & Engine
Q: The speech recognition engine fails to install. What should I do?
This is most likely a network issue. The open-source libraries the engine depends on are hosted on GitHub, which may be unstable to access from networks in mainland China, causing the download to fail. Check your network and try again.
Timestamps & Content Style
Q: Why do the transcribed notes contain timestamps?
The "Key Timestamps" switch is enabled under Settings → Video to Notes. Turn it off to remove timestamps from the notes.
Q: How do I find out when a specific sentence or paragraph appears in the video?
Enable "Key Timestamps" under Settings → Video to Notes, and the note will mark the video time points of key content.
Q: How do I get the original, verbatim transcript?
Set "Content Optimization Level" to "Verbatim" under Settings → Video to Notes to keep the original content without polishing or condensing.
Performance & Acceleration
Q: Speech recognition is very slow and my device gets hot during transcription. What can I do?
Your device may not support GPU acceleration, so recognition falls back to pure CPU mode, which is extremely CPU-intensive. Try switching to a smaller recognition model to reduce the load.
Q: How do I enable GPU acceleration?
GPU acceleration mode works on Windows devices with a CUDA environment, or on Macs with Apple Silicon (M-series) chips via Metal.
Q: My device supports GPU acceleration — why is speech recognition still slow?
There may be compatibility issues with the released build of the third-party engine component whisper.cpp. Try uninstalling the engine tool installed automatically by the app, then follow the manual installation guide to download another version and place it in the specified directory.
Link Import
Q: Why can't video links from Douyin (TikTok) and other platforms be recognized?
Media sniffing relies on third-party components and is limited by the capabilities of the yt-dlp engine, so links from some platforms cannot be parsed. Download the video to your local device by other means first, then import the local file to convert it into a note.