Voice Notes Assistant

Voice Notes Assistant is an Obsidian plugin for recording, transcription, and AI summaries in lecture and meeting workflows.

Documentation Website

• Source page: docs/index.html
• Local preview: open docs/index.html in your browser
• GitHub Pages (recommended):
  1. Repository Settings -> Pages
  2. Source: Deploy from a branch
  3. Branch: main, Folder: /docs
  4. Site URL will be: https://chenqiyukkk.github.io/Voice-Notes-Assistant/

Plugin Metadata

• Plugin ID: lecture-recorder
• Name: Voice Notes Assistant
• Current version: 0.1.2
• Min Obsidian version: 1.0.0
• Desktop only: true

Features

• Inline audio block (lecture-audio) with start/pause/resume/stop controls
• Playback UI with seek, speed control, and waveform
• Transcription providers:
  • OpenAI Whisper-compatible APIs
  • iFLYTEK RAASR
  • Local whisper.cpp (offline)
• Summary providers:
  • OpenAI-compatible chat-completion APIs
  • Claude API (Anthropic)
• Sidecar cache files:
  • *.transcript.json
  • *.summary.md
• UI language switching: Chinese / English

Privacy and External Services

• Local recording files and cache files are stored in your vault by default.
• Cloud requests are sent only when you run transcription/summary with a cloud provider.
• Cloud providers used by this plugin:
  • Whisper-compatible transcription endpoint (/audio/transcriptions)
  • iFLYTEK RAASR API (https://raasr.xfyun.cn/v2/api)
  • OpenAI-compatible chat-completion endpoint (/chat/completions)
  • Claude Messages API (https://api.anthropic.com/v1/messages)
• API keys are stored in plugin settings data (.obsidian/plugins/lecture-recorder/data.json).
• If you choose local whisper.cpp, transcription runs offline on your machine.

Development Install

1. Place this repo in your vault plugins directory (folder name should be lecture-recorder for real installation).
2. Install dependencies and build:

npm install
npm run build

3. In Obsidian: Settings -> Community Plugins -> Reload plugins, then enable Voice Notes Assistant.

Version Management

• Bump version (updates manifest.json, package.json, versions.json together):

npm run version:bump -- 0.1.2

• Bump version and change compatibility floor:

npm run version:bump -- 0.2.0 1.6.0

• Release validation (build + consistency checks):

npm run release:check

Obsidian Community Plugin Submission

Required repository files

• README.md
• LICENSE
• manifest.json
• versions.json

Release assets (GitHub Release)

Each release tag must upload:

• manifest.json
• main.js
• styles.css (if exists)

Submission steps

1. Prepare release files:

npm run release:check

2. Commit and tag:

git add .
git commit -m "chore: prepare 0.1.2 for Obsidian review"
git tag 0.1.2
git push origin main --tags

3. Create a GitHub Release for tag 0.1.2 and upload:

   • manifest.json
   • main.js
   • styles.css

4. Submit to obsidianmd/obsidian-releases:

   • Edit community-plugins.json
   • Add one line:

"https://github.com/chenqiyukkk/Voice-Notes-Assistant"

5. Open PR and wait for review feedback.

License

MIT