Utilities

Self-hosted microservices, local AI inference tools, and media processing utilities.

⚠️ Non-Commercial & User Liability Disclaimer:
Local AI models (OmniVoice, Whisper, Parakeet) operate strictly under non-commercial academic licenses (CC-BY-NC). Any monetization (e.g. YouTube/TikTok revenue, commercial voiceover, paid assets) is strictly prohibited. By accessing these tools, each user accepts full personal legal liability for processed inputs, voice consent, and generated audio under EU DSA Art. 6. Read full legal terms (5 languages) →

VoiceStudio — AI Voice & Speech Lab [Active]
Open-source, local-first ElevenLabs alternative for voice cloning, voice design, text-to-speech, video dubbing, and transcription in 646 languages.
Repository: github.com/debpalash/voicestudio
Neural Audio Demux & Stem Separator [Placeholder / Planned]
Multi-track music source separation (vocals, drums, bass, synths) using Demucs and Whisper audio alignment.
Document OCR & Tabular Data Extractor [Placeholder / Planned]
Vision-language and layout-aware OCR engine to digitize research papers, scanned documents, and tabular financial sheets.
Latent Image Upscaler & Face Restorer [Placeholder / Planned]
Super-resolution diffusion utility for enhancing low-resolution images and reconstructing high-fidelity facial details.
Ephemeral Encrypted Scratchpad [Placeholder / Planned]
Zero-knowledge in-memory pastebin with self-destructing links and automated cryptographic erasure.