https://github.com/razdnut/audiodiary
Personal diary based on vocal notes to be transcribed and synthesized via whisper model and OpenAI Api Key
https://github.com/razdnut/audiodiary
diary diary-application notes openai personal vocal-notes whisper
Last synced: 2 months ago
JSON representation
Personal diary based on vocal notes to be transcribed and synthesized via whisper model and OpenAI Api Key
- Host: GitHub
- URL: https://github.com/razdnut/audiodiary
- Owner: Razdnut
- License: mit
- Created: 2025-09-22T09:28:23.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-09-22T12:01:22.000Z (9 months ago)
- Last Synced: 2025-09-22T12:18:50.924Z (9 months ago)
- Topics: diary, diary-application, notes, openai, personal, vocal-notes, whisper
- Language: TypeScript
- Homepage:
- Size: 436 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Audio Diary (Audio Diario)
A modern, privacy‑friendly journaling web app built with React, TypeScript, Vite, Tailwind, and shadcn/ui. Capture daily notes, record voice memos, transcribe and summarize them with OpenAI, export your data, and track your mood over time.
Key highlights:
- Local‑first: journal data is stored in your browser (localStorage).
- Optional AI: uses your OpenAI API key client‑side for transcription and summaries.
- Mobile‑friendly: tuned header layout and safe‑area padding for Android/iOS.
- i18n: English and Italian with language switching in the UI.
## Goals
The goal is to create an Android app that lets you upload or record an audio file and obtain a transcription using your OpenAI API key and the Whisper model. Afterwards you can optionally generate a summary driven by a customizable prompt when desired.
Manual audio file uploads are now supported alongside in-app recording.
## Features
- Daily notes
- Multiple notes per day (add, select, delete a specific note)
- Rich text area for thoughts and reflections
- 1–5 star rating for the day (with one‑click reset)
- Audio notes and AI
- Record audio in the browser (Web MediaRecorder)
- Upload existing audio files from your device
- Transcribe with Whisper (`whisper-1`) using your OpenAI key
- Summarize transcripts with GPT models (configurable)
- Editable “System Prompt” to customize summaries’ tone and content
- Delete a note’s audio (and all audios at once from Settings)
- Copy either the generated summary or full transcript into your note with one click
- Export
- Export all notes to JSON
- Export to ICS calendar file (one event per note) with localized titles
- Calendar and statistics
- Calendar shows only today highlighted on first load; after you pick a date, days with entries become highlighted
- Stats moved to the bottom of the page: total entries, average rating, number of recordings
- Reset all ratings to zero with one click (does not delete content)
- Internationalization (i18n)
- English and Italian translations
- Dynamic date localization (date‑fns locales)
- Quick language switch in the header and in Settings
- Mobile polish
- Safe‑area insets (top/bottom) for devices with cutouts
- Responsive, centered header — controls fit on small screens
## Screenshots
Add your screenshots here (optional):
- Calendar and daily note view
- Audio controls and AI actions
- Export dialog and Settings
## Getting Started
Prerequisites:
- Node.js 18+ and npm
Install and run:
```bash
npm install
npm run dev
```
Build:
```bash
npm run build
npm run preview
```
## Docker
Two options are provided: run the published image from GitHub Container Registry (GHCR) or build locally.
- Pull and run from GHCR (version 1.0.7):
```bash
docker run --rm -p 8080:80 ghcr.io/razdnut/audiodiary:1.0.7
# Then open http://localhost:8080
```
- Build and run locally:
```bash
docker build -t audiodiary:local .
docker run --rm -p 8080:80 audiodiary:local
```
## Docker Compose
Use the included compose file to build and run:
```bash
docker compose up --build
# Open http://localhost:8080
```
## Android (Capacitor)
Build an Android APK with Capacitor.
Local (Android Studio):
- Prereqs: Android Studio (SDK + build tools), Java 17.
- Required permissions (already declared in `android/app/src/main/AndroidManifest.xml`):
- `RECORD_AUDIO` and `MODIFY_AUDIO_SETTINGS`
- `READ_MEDIA_AUDIO` (Android 13+) and legacy storage permissions for older devices
- Steps:
1. Install deps: `npm install`
2. Build web: `npm run build`
3. Add/sync Android: `npm run cap:add:android` (first time), then `npm run cap:sync`
4. Open in Android Studio: `npm run android:open`
5. Build a Debug APK: from Android Studio, or via Gradle: `cd android && ./gradlew assembleDebug`
6. APK path: `android/app/build/outputs/apk/debug/app-debug.apk`
CI (GitHub Actions):
- This repo includes a workflow that builds a Debug APK on tag push (`v*`) or manual dispatch:
- `.github/workflows/android-apk.yml`
- The APK is uploaded as an artifact and attached to the GitHub Release for the tag.
## OpenAI Setup (Optional)
All AI functionality runs in the browser. Open Settings and paste your OpenAI API key. You can choose:
- Transcription model: `whisper-1`
- Summary model: e.g., `gpt-4o-mini`, `gpt-4o`, `gpt-3.5-turbo`
- System Prompt: customize how summaries are generated
Important: this app uses `dangerouslyAllowBrowser: true` in the OpenAI client. Do not use a production key without a trusted backend. Consider proxying requests in production.
## Data Storage and Privacy
- Notes, ratings, transcripts, summaries, and simple flags are stored in `localStorage`.
- Audio is recorded as a Blob and referenced via a temporary Object URL (blob:). As such, audio blobs are not reliably persisted across page reloads or different sessions. Use Export to save your content or a proper storage backend if you need permanent audio storage.
- You can delete audio per note or delete all audio (and related transcripts/summaries) from Settings.
## Internationalization
- The app supports Italian and English.
- Switch language from the header selector or within Settings.
- Date formatting adapts automatically via `date-fns` locales.
## Exporting
- JSON: Exports all notes as an array (date, content, rating, transcript, summary, audioUrl).
- ICS: Creates calendar events (18:00–18:30 by default) for each note.
- Titles and calendar name are localized.
## Keyboard and UX Tips
- Save Entry: saves the currently selected note for the selected day.
- New note: creates a new blank note for the day and switches to it.
- Delete note: removes just the selected note.
- Clear: clears the current form values (does not delete saved entries until you Save).
## Tech Stack
- React 18 + TypeScript + Vite
- Tailwind CSS + shadcn/ui + Radix UI
- date‑fns for date formatting and locales
- ics for calendar export
## Project Structure
- `src/pages/DailyJournal.tsx` — main journal UI and logic (calendar, notes, stats)
- `src/components/AudioControls.tsx` — recording, transcribing, summarizing, and audio UI
- `src/components/SettingsDialog.tsx` — API key, models, system prompt, language, delete all audio
- `src/components/ExportDialog.tsx` — export to JSON/ICS
- `src/i18n/i18n.tsx` — i18n provider, translations, and date locales
- `src/services/openai.ts` — OpenAI calls (transcription and summaries)
- `src/utils/export-utils.ts` — JSON/ICS export helpers
## Roadmap Ideas
- Persistent audio storage (IndexedDB or backend upload)
- Per‑note tags and search
- Charts and insights for mood tracking over time
## License
MIT — see [LICENSE](./LICENSE) for details.