KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#dsh #dsh-plugin #deepseek-harness #typescript #ai-agent #codex #developer-tools #claude-code #windows #mcp #python #desktop-app

AIAudioTranscription

★ 8 Open GitHub ↗

Simple Android app for transcribing voice messages and media files to text using OpenAI's Whisper API. Features encrypted API key storage, local history, and AI-powered text cleanup for improved readability. Perfect for converting audio messages from various messenger apps into readable text.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Android AI Transcription App

[GitHub Repository](https://github.com/dhcgn/AIAudioTranscription) • [Copilot Instructions](.github/copilot-instructions.md)

## Overview

This app lets you easily transcribe media files (audio or video) using AI. Users can upload, share, or open media files which are then re-encoded and sent to multiple transcription models. The app supports traditional Whisper-1 transcription, as well as GPT-4o Transcribe and GPT-4o-mini Transcribe models. AI-powered post-processing further improves transcript readability. All transcriptions and settings are stored locally—with your API key secured in encrypted storage.

## Key Features

- **Media Handling:**  
  - Select, share, or open media files directly from your device.
  - Direct transcription of voice messages from messaging apps like WhatsApp, Telegram, or Signal.
  - Automatic re-encoding using **Media3 Transformer** to **M4A/AAC** format (adaptive bitrate 16–32 kbps) for all transcription models.
  - File size check (max. 24MB after processing) to ensure smooth operation.

- **Multiple Transcription Models:**  
  - **Whisper-1:** Traditional audio transcription model.
  - **GPT-4o Transcribe:** Specifically optimized for transcription tasks.
  - **GPT-4o-mini Transcribe:** Efficient transcription model for cost-effective processing.

- **AI-Powered Cleanup:**  
  - Enhance transcript readability with AI-driven cleanup using OpenAI chat completions (GPT-5-nano).
  - Customizable cleanup prompts ensure the original content is preserved while improving clarity.
  - Optional auto-format setting to automatically enhance readability after every transcription.

- **Local History & Settings:**  
  - Maintain a local history of transcriptions with comprehensive statistics:
    - Transcript length (character count)
    - File sizes (original and uploaded)
    - Audio duration (when available)
    - Processing settings used (model, language, prompt)
  - History view displays only relevant information for each entry, with expandable details.
  - Secure API key storage using EncryptedSharedPreferences.
  - In-app Settings allow you to:  
    - Save and test your OpenAI API key.  
    - Configure transcription models and language preferences.  
    - Customize prompts for both transcription and AI cleanup.
    - Enable auto-format to automatically enhance transcriptions after processing.

- **User Experience Enhancements:**  
  - Support for shared intents (from other apps) and direct file access.
  - Retry functionality for reprocessing files with updated parameters.
  - Clear, responsive UI built with Jetpack Compose and modern Android architecture practices.

- **Use Cases:**
  - Convert voice messages to text for easy reading and sharing
  - Archive and search through voice message content
  - Make voice messages accessible for hearing-impaired users
  - Quick transcription of meeting recordings and lectures

## How It Works

1. **Initialization & Setup:**  
   - Enter and test your OpenAI API key in the Settings screen.
   - Choose your transcription model and set any custom prompts or language preferences.

2. **Media Processing:**  
   - Select a media file, or share one to the app.
   - The file is copied, re-encoded to AAC format, and its size is validated.

3. **Transcription & Cleanup:**  
   - The processed file is uploaded to the selected transcription API.
   - Once transcribed, the text is optionally enhanced with an AI cleanup process.
   - The final transcript is displayed and stored locally with comprehensive statistics:
     - File sizes (original and processed)
     - Transcript length
     - All processing parameters used

4. **History & Management:**  
   - View past transcriptions with detailed statistics (file sizes, transcript length, duration).
   - Each entry shows only relevant processing settings that were actually used.
   - Copy transcriptions with or without detailed statistics.
   - Delete individual entries or clear all history as needed.

## Flowchart

```mermaid
graph TD
    A[Start] --> B[Enter & Test API Key in Settings]
    B --> C[Select Transcription Model & Set Prompts]
    C --> D[Choose Media File / Share to App]
    D --> E["Copy & Re-encode File (M4A/AAC)"]
    E --> F["File Size Check (<= 24MB)"]
    F --> G[Upload to Selected Transcription API]
    G --> H{Response Successful?}
    H -- Yes --> I[Display & Store Transcript]
    I --> J[Optional: AI Cleanup for Readability]
    J --> K[Update History]
    H -- No --> L[Show Error Message]
    K --> M[User Can Retry or View History]
    M --> N[End]
```

## Privacy & Legal Notice

- **Third-Party Processing:**  
  Your media files are sent to OpenAI’s servers. Ensure you have permission to share and transcribe them.
  
- **Sensitive Data:**  
  Do not transcribe content with sensitive personal or confidential information.

- **User Responsibility:**  
  Use your own OpenAI API key. You are responsible for any costs incurred, and you must comply with OpenAI’s Terms of Service.

## Testing

For developers and testers who want to build the app with a pre-configured OpenAI API key for testing, see the [Testing Guide](TESTING.md). This allows you to create builds with embedded API keys without committing secrets to the repository.

## Next Steps
Future enhancements include:

- Refactoring UI state management into ViewModels.
- Consolidating duplicated code and further adopting Kotlin coroutines with Retrofit’s suspend functions.
- Enhancing the UI and user interaction flow.

AIAudioTranscription

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README