Voice To Text Python App

.claude/settings.local.json

@@ -0,0 +1,15 @@
+{
+  "permissions": {
+    "allow": [
+      "Bash(python -m pip:*)",
+      "Bash(where python)",
+      "Bash(python:*)",
+      "Read(//c/Users/Dan Zambello/.claude/agents/**)",
+      "Read(//c/Users/Dan Zambello/.claude/**)",
+      "mcp__browsermcp__browser_snapshot",
+      "mcp__browsermcp__browser_click"
+    ],
+    "deny": [],
+    "ask": []
+  }
+}

CLAUDE.md

@@ -0,0 +1,76 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+This is OpenAI's Whisper repository - a general-purpose speech recognition model that performs multilingual speech recognition, speech translation, and language identification. The codebase is built as a Python package with PyTorch.
+
+## Architecture
+
+### Core Components
+
+- **whisper/__init__.py**: Main entry point with model loading (`load_model()`) and available models registry
+- **whisper/model.py**: Core Whisper transformer model implementation
+- **whisper/transcribe.py**: High-level transcription interface with CLI entry point
+- **whisper/decoding.py**: Lower-level decoding logic and options
+- **whisper/audio.py**: Audio processing utilities (loading, mel spectrograms, padding)
+- **whisper/tokenizer.py**: Text tokenization and language handling
+- **whisper/normalizers/**: Text normalization for different languages
+
+### Model Architecture
+- Transformer sequence-to-sequence model
+- Multiple model sizes: tiny, base, small, medium, large, turbo
+- Both English-only (.en) and multilingual variants
+- Models downloaded from Azure CDN and cached locally
+
+## Development Commands
+
+### Testing
+```bash
+pytest                    # Run all tests
+pytest tests/test_*.py    # Run specific test file
+pytest -m requires_cuda   # Run CUDA-specific tests
+```
+
+### Code Quality
+```bash
+black .                   # Format code
+isort .                   # Sort imports
+flake8                    # Lint code
+pre-commit run --all-files # Run all pre-commit hooks
+```
+
+### Installation for Development
+```bash
+pip install -e .[dev]     # Install in development mode with dev dependencies
+```
+
+## Package Structure
+
+- Built using setuptools with pyproject.toml configuration
+- Entry point: `whisper` command maps to `whisper.transcribe:cli`
+- Dependencies: torch, numpy, tiktoken, tqdm, numba, more-itertools
+- Optional triton dependency for Linux x86_64 optimization
+
+## Key APIs
+
+### High-level Usage
+```python
+import whisper
+model = whisper.load_model("turbo")
+result = model.transcribe("audio.mp3")
+```
+
+### Lower-level Usage
+```python
+audio = whisper.load_audio("audio.mp3")
+mel = whisper.log_mel_spectrogram(audio)
+result = whisper.decode(model, mel, options)
+```
+
+## Testing Notes
+
+- Tests use pytest with custom markers for CUDA requirements
+- Random seeds fixed for reproducibility (seed=42)
+- Test coverage includes audio processing, normalization, timing, tokenization, and transcription

README_SETUP.md

@@ -0,0 +1,137 @@
+# Voice to Text Converter - Setup Guide
+
+## Quick Start
+
+### Prerequisites
+- Windows 10/11
+- Internet connection (for initial setup)
+- Microphone access
+
+### Installation Steps
+
+1. **Install Python** (if not already installed)
+   - Download from [python.org](https://python.org)
+   - **IMPORTANT**: Check "Add Python to PATH" during installation
+   - Minimum version: Python 3.8
+
+2. **Run Setup**
+   - Double-click `setup.bat`
+   - Wait for dependencies to install (may take 5-10 minutes)
+   - Setup is complete when you see "Setup Complete!"
+
+3. **Start Using**
+   - **GUI Mode**: Double-click `voice_to_text_gui.bat`
+   - **Terminal Mode**: Double-click `voice_to_text_terminal.bat`
+
+## Usage Modes
+
+### GUI Mode (Recommended)
+- **Launch**: `voice_to_text_gui.bat` (batch window closes, GUI stays open)
+- **Features**:
+  - Visual interface with buttons
+  - F1 global hotkey for recording
+  - Always on top option
+  - System tray integration
+  - Settings dialog
+- **Best for**: Daily use and continuous workflow
+
+### Terminal Mode
+- **Launch**: `voice_to_text_terminal.bat`
+- **Features**:
+  - Simple text interface
+  - Press Enter to stop recording
+  - Lightweight and fast
+- **Best for**: Quick one-off recordings
+
+## Creating Desktop Shortcuts
+
+1. **Right-click** on desktop → **New** → **Shortcut**
+2. **Browse** to the batch file you want (e.g., `voice_to_text_gui.bat`)
+3. **Name** the shortcut (e.g., "Voice to Text")
+4. **Optional**: Right-click shortcut → **Properties** → **Change Icon**
+
+## First Run
+
+- **Model Download**: First run will download Whisper model (~150MB)
+- **Microphone Permission**: Windows may ask for microphone access
+- **Settings**: GUI mode creates `voice_to_text_settings.json` for preferences
+
+## File Structure
+
+```
+voice-to-text/
+├── voice_to_text.py           # Main application
+├── setup.bat                  # One-time setup
+├── voice_to_text_gui.bat      # GUI launcher
+├── voice_to_text_terminal.bat # Terminal launcher
+├── requirements.txt           # Python dependencies
+├── transcripts/               # Saved transcriptions
+├── voice_to_text_settings.json # Settings (created after first GUI run)
+└── README_SETUP.md           # This file
+```
+
+## Voice Commands
+
+The system automatically converts natural speech into Claude Code prompts:
+
+| Say This | Gets Converted To |
+|----------|-------------------|
+| "use agent python pro" | `@agent python-pro` |
+| "run tool bash" | `@tool bash` |
+| "file package.json" | `@file package.json` |
+| "directory source" | `@dir source/` |
+| "function get user" | `` `getUser()` function`` |
+
+## Troubleshooting
+
+### "Python is not installed"
+- Install Python from [python.org](https://python.org)
+- **Must check "Add Python to PATH"** during installation
+- Restart command prompt/computer after installation
+
+### "Failed to install dependencies"
+- Check internet connection
+- Try running `setup.bat` as administrator
+- Manually run: `pip install -r requirements.txt`
+
+### "No audio recorded"
+- Check microphone permissions in Windows Settings
+- Ensure microphone is not muted
+- Try a different microphone
+
+### "Poor transcription accuracy"
+- Speak clearly and at normal pace
+- Reduce background noise
+- Move closer to microphone
+- In GUI mode: Settings → Change to larger Whisper model
+
+### GUI Hotkey Not Working
+- Check if another application is using F1
+- Try running as administrator
+- Change hotkey in Settings dialog
+
+### System Tray Issues
+- If tray doesn't work, app falls back to normal minimize
+- Some Windows configurations don't support system tray
+- This doesn't affect core functionality
+
+## Advanced Settings (GUI Mode)
+
+Access via Settings button:
+- **Global Hotkey**: Change from F1 to F2-F12
+- **Whisper Model**: tiny (fast) to large (accurate)
+- **Always on Top**: Keep window visible
+- **Auto Copy**: Automatically copy to clipboard
+
+## Support
+
+- Check transcripts in `/transcripts` folder
+- Settings saved in `voice_to_text_settings.json`
+- For issues, check the console output in terminal mode
+
+## Updates
+
+To update the application:
+1. Replace `voice_to_text.py` with new version
+2. Update `requirements.txt` if needed
+3. Run `setup.bat` again if dependencies changed

assessment_analysis.json

@@ -0,0 +1,153 @@
+{
+  "Q1": {
+    "question": "What is the purpose of determining and documenting requirements for a cabinet installation?",
+    "transcribed_answer": "To complete it accurately and efficiently, including detailed specification for the cabinets and all components.",
+    "written_summary": "To have a better understanding of the project for all persons involved and plan in advance for any difficulties and refer to documents as a guide for future projects.",
+    "assessment": {
+      "word_count": 15,
+      "has_substantial_content": true,
+      "keyword_relevance": 0.2857142857142857,
+      "transcription_summary_match": 0.15151515151515152
+    },
+    "transcription_details": {
+      "text": "To complete it accurately and efficiently, including detailed specification for the cabinets and all components.",
+      "language": "en",
+      "segments": [
+        {
+          "id": 0,
+          "seek": 0,
+          "start": 0.0,
+          "end": 8.0,
+          "text": " To complete it accurately and efficiently, including detailed specification for the cabinets and all components.",
+          "tokens": [
+            50364,
+            1407,
+            3566,
+            309,
+            20095,
+            293,
+            19621,
+            11,
+            3009,
+            9942,
+            31256,
+            337,
+            264,
+            37427,
+            293,
+            439,
+            6677,
+            13,
+            50764
+          ],
+          "temperature": 0.0,
+          "avg_logprob": -0.3367410898208618,
+          "compression_ratio": 1.1914893617021276,
+          "no_speech_prob": 0.07899459451436996
+        }
+      ]
+    }
+  },
+  "Q2": {
+    "error": "File not found: 77914809189__571E73A4-D2E8-4B00-934C-5B2E54DE47A4.MOV"
+  },
+  "Q3": {
+    "question": "What information is found in the appliance manuals?",
+    "transcribed_answer": "It's including product identifications, safety warnings and precautions step by step operating instructions, installation and assembly instructions. My tenants got lined, troubleshooting tips, technical specifications and warranty information.",
+    "written_summary": "Fitting instructions and requirements",
+    "assessment": {
+      "word_count": 28,
+      "has_substantial_content": true,
+      "keyword_relevance": 0.25,
+      "transcription_summary_match": 0.03571428571428571
+    },
+    "transcription_details": {
+      "text": "It's including product identifications, safety warnings and precautions step by step operating instructions, installation and assembly instructions. My tenants got lined, troubleshooting tips, technical specifications and warranty information.",
+      "language": "en",
+      "segments": [
+        {
+          "id": 0,
+          "seek": 0,
+          "start": 0.0,
+          "end": 7.0,
+          "text": " It's including product identifications, safety warnings and precautions",
+          "tokens": [
+            50364,
+            467,
+            311,
+            3009,
+            1674,
+            2473,
+            7833,
+            11,
+            4514,
+            30009,
+            293,
+            34684,
+            50714
+          ],
+          "temperature": 0.0,
+          "avg_logprob": -0.3135071884502064,
+          "compression_ratio": 1.5576923076923077,
+          "no_speech_prob": 0.02320166677236557
+        },
+        {
+          "id": 1,
+          "seek": 0,
+          "start": 7.0,
+          "end": 13.0,
+          "text": " step by step operating instructions, installation and assembly instructions.",
+          "tokens": [
+            50714,
+            1823,
+            538,
+            1823,
+            7447,
+            9415,
+            11,
+            13260,
+            293,
+            12103,
+            9415,
+            13,
+            51014
+          ],
+          "temperature": 0.0,
+          "avg_logprob": -0.3135071884502064,
+          "compression_ratio": 1.5576923076923077,
+          "no_speech_prob": 0.02320166677236557
+        },
+        {
+          "id": 2,
+          "seek": 0,
+          "start": 13.0,
+          "end": 21.0,
+          "text": " My tenants got lined, troubleshooting tips, technical specifications and warranty information.",
+          "tokens": [
+            51014,
+            1222,
+            31216,
+            658,
+            17189,
+            11,
+            15379,
+            47011,
+            6082,
+            11,
+            6191,
+            29448,
+            293,
+            26852,
+            1589,
+            13,
+            51414
+          ],
+          "temperature": 0.0,
+          "avg_logprob": -0.3135071884502064,
+          "compression_ratio": 1.5576923076923077,
+          "no_speech_prob": 0.02320166677236557
+        }
+      ]
+    }
+  }
+}