The Obsidian Transcriber plugin converts an audio file or YouTube video into an Obsidian note. Both the metadata and the text are captured in the note.
YouTube video: Bluelab Pulse Meter Review
After the plugin finishes its work with the Obsidian Transcriber Service, the created note includes:
- YouTube metadata as YAML frontmatter.
- A YouTube URL "button" that when clicked invokes a YouTube player in Obsidian's sidebar.
- 👍(Assuming the Timestamp Notes community plugin is installed and enabled)
- The transcript broken into chapters with topic titles and timestamps.
- Below each topic title is a timestamp "button" that when clicked will move the player to that time in the video.
- 👍(Assuming the Timestamp Notes community plugin is installed and enabled)
YouTube video: Jeff Lowenfels - The Updated Soil Food Web
The features are identical to the previous example except instead of breaking the text into chapters, the transcribed text is broken up into time segments.
Compared with audio sources that originate from YouTube, the front matter of an uploaded audio file contains far less metadata. The transcript is broken into chapters based on the time chunk setting in the settings UI.
- Go to your Obsidian vault's plugin folder (
<vault name>/.obsidian/plugins
). - Create the folder
obsidian-transcriber-plugin
. - Go to release.
- Copy main.js and manifest.json to the
obsidian-transcriber-plugin
folder. - Open your Obsidian vault and go into Settings. Choose core plugins and enable the
Obsidian Transcriber Plugin
.
Settings
> Community plugins
> Browse
.Install
.Community plugins
section.Click on the ❀ flower in the left ribbon to bring up the UI
- OR -
<ctrl-p>
(PC) <cmd-p>
(Mac) then type t
then r
.
Enter either a YouTube URL, or click on the button to choose an audio file. Then click the Submit
button.
Progress notifications will start to be shown. It can take quite a bit of time to complete a transcription, particularly if the desired audio quality loaded a larger model.
You can configure the plugin settings by navigating to Settings
> Plugin Options
> Obsidian Transcriber
.
The default values are set in plugin_settings.ts
- 📂 Transcripts Folder: Specify the vault folder for transcripts. The folder will be created if it does not exist.
- 🌐 Obsidian Transcriber Service URL: Set the URL to the endpoint running the Obsidian Transcriber Service endpoint.
- 🎧 Audio Quality: Choose audio quality from
tiny
tolarge
. Higher quality increases processing time. - 💻 Compute Type: The compute type determines the precision and efficiency of the computations performed by the model.
- 📘 Chapter Chunk Time: The amount of time to chunk the transcript into chapters if there is no chapter info in the metadata.
- 🐞 Debug: If toggled on, debug, warning, and error messages are sent to the console debugger. If not on, only error messages are sent to the console.
To troubleshoot, set the logging to debug and check the console for messages. The steps:
- Toggle on the debug setting in the settings UI.
- Open the console debugger.
- On Windows, press
Ctrl+Shift+I
. - On Mac, press
Cmd+Option+I
.
- On Windows, press
- Press
Ctrl-L
(or equivalent on Mac) to clear the console. - Run the plugin.
Debug, warning, and error messages will be sent to the console. If you cannot fix the issue, please:
- Save the console output by right-clicking on the console and choosing
Save as...
.
- Open an issue and attach the saved console output.