FreeFlow Studio 🚀

Next-Gen AI Voice Dictation, Voiceover QA Validator & Acoustic Prosody Analyzer for Windows.

FreeFlow Studio is a high-performance Windows desktop suite engineered for video creators, crypto marketers, podcasters, and AI voice producers. It combines ultra-fast global voice dictation with automated STT voiceover validation, acoustic human-likeness scoring (Prosody Analysis), and batch audio conversion.

---

📑 Table of Contents

• 🚀 Quick Start
• 🌟 Key Features
• 💻 Offline Local Whisper AI
• 🔒 Security & Key Protection
• 🛡️ Antivirus & SmartScreen Notes
• 🏗️ Architecture & Technology Stack
• 🛠️ Build & Installation Guide
• ⚖️ Third-Party Libraries & Licensing
• 📜 License

---

🚀 Quick Start

Option A: Pre-built Binary Release (Recommended)

1. Download Release: Head to GitHub Releases and download FreeFlowStudio-win-x64.zip.
2. Extract & Launch: Unpack the ZIP archive and run FreeFlowWin.App.exe.
3. Start Dictating: Hold [F9] (or [Ctrl+Space]) to speak, release to instantly transcribe and paste into your active window!

Option B: Run from Source (.NET 8.0 SDK)

git clone https://github.com/bbimer/whisper-freeflow-studio.git
cd whisper-freeflow-studio
dotnet run --project FreeFlowWin.App/FreeFlowWin.App.csproj -c Release

---

🌟 Key Features

🎙️ 1. Global AI Voice Dictation

• System-Wide Hotkey Activation: Trigger dictation instantly from any app using hardware-level keyboard hooks with multi-modifier release tracking (Ctrl + Space, F9).
• Groq Whisper STT Speed: Powered by whisper-large-v3-turbo for near-instant speech recognition (~0.35s) with custom domain vocabulary prompts.
• Dual Processing Modes:
  • ⚡ Instant Mode (~0.3s): Ultra-low latency voice-to-text directly from Whisper Turbo.
  • 🧠 LLM Smart Mode (~3s): Deep AI post-processing (grammar correction, filler-word removal, and active window context alignment).
• Proxy Tunnel & Anti-Block Gateway: Configurable proxy endpoint and toggle to bypass Cloudflare/Groq regional 403 Forbidden errors without system-wide VPNs.
• Auto-Paste: Direct text injection into active fields, IDEs, text editors, or web browsers.

🎯 2. QA Voiceover & STT Script Validator

• Batch Drag & Drop: Drop dozens of audio files (.mp3, .wav, .m4a, .flac) simultaneously for bulk QA testing.
• Word-by-Word Diff Matrix: Visual color-coded token badges displaying exact text matches, mispronunciations, prompter deletions, and extra insertions.
• Per-File Report Cards: Dedicated breakdown cards for every audio take to easily compare performance.

---

🧠 3. Acoustic Prosody & AI Naturalness Detector

• Human-Likeness Score (NATURALNESS %): Calculates RMS energy variance, micro-pause frequency, and pitch dynamics to distinguish organic human recordings from robotic AI audio.
• Audience Protection: Flags robotic monotonic voiceovers before publishing to Tier-1 traffic channels.
• Actionable AI Feedback: Recommends exact adjustment parameters for ElevenLabs voice generation.

---

🎚️ 4. ElevenLabs Voice Optimization & Syntax Hacks

• Preset Cheat Sheet: Recommended stability (35–45%), similarity (75–85%), and style exaggeration settings for realistic institutional voiceovers.
• One-Click Humanizer: Automatically injects breath dashes (—), pauses (...), and organic filler starters (Look,, Well,) into target scripts.

---

⚡ 5. Batch Media & Audio Converter

• Multi-File Processing: Convert audio and video files (.mp3, .wav, .m4a, .flac, .mp4, .mov, .avi) in parallel.
• Audio Extraction: Rip high-fidelity audio streams directly from video footage.
• Custom DSP Parameters: Adjustable sample rates (44.1kHz, 48kHz), bitrates (up to 320 kbps), and mono/stereo channels.

---

📊 6. Telemetry & Productivity Tracking

• Tracks daily, weekly, and monthly dictated word counts.
• Displays voice-to-typing speedup factor (e.g. 3.8x faster) and total hours saved.
• 100% Local Storage: All stats are stored strictly on your local PC (%AppData%\FreeFlowWindows\stats.json). See PRIVACY.md.

---

💻 Offline Local Whisper AI (Zero VPN / Zero API)

FreeFlow Studio features a 100% standalone offline transcription engine powered by Whisper.net and ggml C++ runtime.

Why Use Local Mode?

• 🔒 Complete Privacy: Audio never leaves your computer.
• 🌐 Bypass VPN / Network Blocks: Zero dependence on cloud APIs or Cloudflare WAF restrictions.
• 💸 Unlimited & Free: No API quotas, usage limits, or subscription costs.

Available GGML Models & SHA256 Integrity Hashes

Model	File Size	VRAM / RAM	Speed	SHA256 Checksum	Best For
ggml-base.bin	~140 MB	~500 MB	⚡⚡⚡ Instant	60ed5bc226b64f19985ea1053e3047b21ac70d5df564c7ebd00b48f07bd546f8	Quick dictation & low-spec laptops
ggml-small.bin (Default)	~460 MB	~1.0 GB	⚡⚡ Very Fast	1be3a9b2063867b937e64e2ec7483364a79917e157fa98c5d94b5c1fffea987b	Balanced daily dictation (Russian & English)
ggml-large-v3-turbo.bin	~1.5 GB	~3.0 GB	⚡ Fast	5639644d6735a4d651a2d5e381023730e666a0d0a2185b306b4d375084931a54	Technical terms, accents, & complex scripts

Verification Procedure

The application automatically verifies the SHA256 checksum of downloaded model files prior to loading into memory. You can manually verify model file integrity in PowerShell:

Get-FileHash "$env:APPDATA\FreeFlowWindows\Models\ggml-small.bin" -Algorithm SHA256

---

🔒 Security & Key Protection

• Windows DPAPI Encryption: Cloud API keys (Groq, ElevenLabs) stored in %AppData%\FreeFlowWindows\settings.json are encrypted using Windows Data Protection API (ProtectedData.Protect scoped to CurrentUser).
• Environment Variables: API keys can also be supplied via GROQ_API_KEY and ELEVENLABS_API_KEY environment variables.
• For full security disclosures and vulnerability reporting policies, see SECURITY.md and PRIVACY.md.

---

🛡️ Antivirus & SmartScreen Notes

FreeFlow Studio uses low-level Windows API hooks (SetWindowsHookEx and SendInput) to capture keypresses globally and simulate text pasting into active applications.

Note

Some antivirus software or Windows SmartScreen may issue a warning for unsigned executables using global hooks. This is a false positive. Official release binaries are signed with a digital code-signing certificate. If compiling locally, you can safely add the build directory to your antivirus exclusions.

---

🏗️ Architecture & Technology Stack

• Framework: C# / .NET 8 WPF
• Design System: WPF.Ui (Fluent Dark Mode theme with native Windows 11 glassmorphism)
• STT Engine: Groq Cloud REST API (whisper-large-v3-turbo) & Whisper.net (Offline C++ runtime)
• Audio DSP: NAudio & MediaFoundation API
• System Hooks: Windows P/Invoke (SetWindowsHookEx, SendInput, GetAsyncKeyState)

FreeFlowWin.slnx
 ├── FreeFlowWin.App    (WPF Client Application, XAML views & UI controllers)
 ├── FreeFlowWin.Core   (STT API client, Prosody Analyzer, Audio Engine, Settings)
 └── FreeFlowWin.TestApp(Automated test harness for core audio & QA modules)

---

🛠️ Build & Installation Guide

Prerequisites

• Operating System: Windows 10 (1903+) or Windows 11
• Runtime: .NET 8.0 SDK

Building from Source

# Clone the repository
git clone https://github.com/bbimer/whisper-freeflow-studio.git
cd whisper-freeflow-studio

# Restore and build the solution
dotnet build FreeFlowWin.slnx -c Release

# Publish self-contained executable
dotnet publish FreeFlowWin.App/FreeFlowWin.App.csproj -c Release -r win-x64 --self-contained false -o ./publish

---

⚖️ Third-Party Libraries & Licensing

FreeFlow Studio incorporates open-source components and cloud API integrations:

Component	License / Terms	Commercial Usage	Description / Terms Link
Whisper.net	MIT License	Permitted	.NET binding for whisper.cpp offline engine
WPF.Ui	MIT License	Permitted	Modern Fluent UI design controls for WPF
NAudio	MIT License	Permitted	Audio recording, playback & DSP framework
Groq Cloud API	Groq Terms	Subject to API ToS	Fast Cloud Speech-to-Text inference engine
ElevenLabs API	ElevenLabs Terms	Subject to Tier ToS	Voice generation preset & prosody optimization

---

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.