An intelligent brainstorming assistant that listens to conversations, analyzes discussions, and contributes ideas using advanced AI models (Claude Sonnet 4.5, GPT-4, Gemini 2.0).

• Multi-Platform Support: Available as mobile apps (iOS/Android) and desktop applications (Windows/macOS/Linux)
• Voice Recognition: Real-time speech-to-text with speaker diarization
• Voice Calibration & Mapping: Optional per-session calibration step where each participant briefly introduces themselves so diarized speaker tags can be mapped to their first names
• AI Models: Support for Claude Sonnet 4.5, GPT-4, and Gemini 2.0 Pro
• Smart Contributions: AI analyzes conversations, contributes relevant ideas, and targets a length close to the average number of words used by human participants
• Configurable Frequency: Tune how often the AI participates
• Voice Selection: Choose between male and female AI voices
• Session Reports: Comprehensive summaries with full transcripts and AI-generated, Word-compatible (RTF) reports saved locally
• Real-time Updates: WebSocket-based real-time session synchronization
• Session-Level Recording: Optional full-session audio recording stored locally on the device under an aibro/audios folder (Linux, Windows, macOS, Android, iOS)
• Previous Sessions Library: Browse, play, share, or delete past session recordings from the device

• Spring Boot 3.2.1 with Java 17
• RESTful APIs for session management
• WebSocket support for real-time communication
• Integration with Claude, OpenAI, and Gemini APIs
• Google Cloud Speech-to-Text and Text-to-Speech
• H2 in-memory database (easily switchable to PostgreSQL)

• Cross-platform mobile and desktop application
• Provider/BLoC for state management
• Audio recording and playback
• Real-time WebSocket communication
• Beautiful, responsive UI

For a more detailed, step-by-step setup guide, see QUICKSTART.md.

• Java 17 or higher
• Maven 3.6+
• Flutter 3.0+
• API Keys:
  • Anthropic (Claude)
  • OpenAI (GPT-4)
  • Google Cloud (Gemini & Speech services)

1. Navigate to the backend directory:

cd backend

2. Build the project:

mvn clean install

3. Run the application:

mvn spring-boot:run

The backend will start on http://localhost:8080

1. Navigate to the Flutter app directory:

cd flutter_app

2. Install dependencies:

flutter pub get

3. Run on your preferred platform:

# For Android
flutter run -d android

# For iOS
flutter run -d ios

# For Desktop (Windows)
flutter run -d windows

# For Desktop (macOS)
flutter run -d macos

# For Desktop (Linux)
flutter run -d linux

Users must configure their API keys in the app settings:

1. Open the app
2. Navigate to Settings
3. Enter your API keys:
   • Claude API Key
   • OpenAI API Key
   • Google Cloud API Key (for Speech services and Gemini)

1. Start a Session:

   • First, configure AI model, contribution frequency (1–20), and AI voice (male/female) in the Settings screen
   • From the Home screen, tap Start Session
   • Enter the number of participants (integer between 1 and 6)
   • Enter each participant's first name
   • Define the objective of the brainstorming session
   • If there are at least 2 participants, you can optionally run a short Calibrate voices step where, in order, each participant briefly introduces themselves so AiBro can map diarized voices to their first names
   • During the session, you can optionally start a session-level recording from the session header card; the audio will be stored locally on the device and can be reviewed later

2. During the Session:

   • The app listens to conversations
   • AI analyzes discussions in real-time
   • When AI has an idea:
     • Blue light indicator appears
     • Sound notification plays
     • AI speaks its contribution
   • AI contributions aim for a word count similar to the average length of human contributions, so the assistant does not dominate the conversation
   • You can toggle a session timer from the session header card, either as a simple elapsed-time indicator or as a countdown to a target session duration; when running, the timer is prominently displayed above the session controls

3. End Session:

   • Tap End session in the session header; a confirmation dialog will appear before closing the session
   • If there is no discussion and no full-session recording, you are returned to the Home screen
   • If there is either a discussion or a full-session recording, you are taken to the Report screen for that session
   • From the Report screen you can generate a comprehensive, Word-compatible (RTF) report that includes participants, objective, selected AI model, duration (when available), ideas grouped by participant, a contributions timeline, and an AI-refined textual summary

4. Review Previous Sessions (Recordings):

   • From the Home screen, tap Previous sessions
   • See a list of all full-session audio recordings stored on the current device, with session id, objective, timestamp, and file name
   • For each recording you can:
     • Play or stop the audio directly in the app
     • Open the recording in the system's default file handler (or its containing folder on desktop platforms)
     • Export/share the audio via email or other apps
     • Delete the recording from the device

5. Local File Storage (Reports & Audio):

   • On each platform, AiBro uses the OS documents directory as a base and creates an aibro folder
   • Reports are generated as Word-compatible .rtf files under aibro/reports
   • Full-session audio recordings are stored under aibro/audios

• POST /api/sessions - Create new session
• GET /api/sessions - List all sessions
• GET /api/sessions/{id} - Get session details
• PUT /api/sessions/{id} - Update session
• POST /api/sessions/{id}/end - End session
• GET /api/sessions/{id}/report - Get session report

• POST /api/audio/transcribe - Transcribe audio with speaker diarization
• POST /api/audio/synthesize - Convert text to speech
• POST /api/audio/calibrate - Calibrate diarized speakers against participant first names using a short introduction round

• POST /api/keys - Save API keys
• GET /api/keys/validate/{model} - Validate API key for specific model

• /ws - WebSocket endpoint
• /topic/session/{id}/ai - AI contributions
• /topic/session/{id}/status - Session status updates
• /topic/session/{id}/contributions - Real-time contributions

aibro/
├── backend/                    # Spring Boot backend
│   ├── src/main/java/com/aibro/
│   │   ├── config/            # Configuration classes
│   │   ├── controller/        # REST & WebSocket controllers
│   │   ├── dto/               # Data transfer objects
│   │   ├── model/             # JPA entities
│   │   ├── repository/        # Data repositories
│   │   └── service/           # Business logic
│   ├── src/main/resources/
│   │   └── application.yml    # Application configuration
│   └── pom.xml
├── flutter_app/               # Flutter frontend
│   ├── lib/
│   │   ├── models/           # Data models
│   │   ├── services/         # API & business services
│   │   ├── screens/          # UI screens
│   │   ├── widgets/          # Reusable widgets
│   │   └── utils/            # Utility functions
│   ├── assets/               # Images, sounds, animations
│   └── pubspec.yaml
└── README.md

• Spring Boot 3.2.1
• Spring Data JPA
• Spring WebSocket
• Spring Security
• OkHttp3
• H2 Database
• Lombok
• Maven

• Flutter 3.0+
• Provider (State Management)
• Dio (HTTP Client)
• WebSocket Channel
• Audio Recorder
• Audio Players
• PDF / report export (including Word-compatible RTF generation stored under aibro/reports)
• Shared Preferences

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

This project is licensed under the MIT License.

For issues and questions, please create an issue in the repository.