Logseq Transcriber Plugin

Quickly transcribe audio files embedded in your Logseq notes using the Deepgram API. This plugin adds a “Transcribe” button next to audio players, allowing you to get a text version of your audio content inserted directly as a new block. If you like it, please give it a star on Github !

If you really like it, you can ☕ buy me a coffee ❤️.

Logseq.mp4

Features

• Inline Transcription: Adds a “🎙️ Transcribe” button next to audio players in your Logseq blocks.
• Configurable: Set your Deepgram API key, preferred transcription model, and audio language via plugin settings.
• Automatic Language Detection: If no language is specified, the plugin will ask Deepgram to auto-detect the language.
• Simple Workflow: Click, transcribe, and get the text as a new block below the audio.

Setup & Configuration

Download the latest release from the plugin marketplace.

Configure Settings

Before using, you'll need to configure the plugin settings.

• Go to the plugin settings:
  • API Key: Your Deepgram API Key (from the Deepgram Console)
  • Transcription Model: e.g. nova-3, whisper-large (defaults to nova-3)
  • Audio Language: Choose a language or “Auto-detect” (see Deepgram Languages)

Usage

1. Embed an audio file in a Logseq block
2. A “🎙️ Transcribe” button appears next to the audio player
3. Click it—the text changes to “🎙️…” while processing
4. The transcription is inserted as a new block beneath the audio

Notes

• Requires an internet connection and a valid Deepgram key with credits
• Accuracy depends on audio quality and chosen model
• Audio is fetched from Logseq’s local asset URL (http://localhost:…/assets/…)

License

MIT