Pure Rust Open-Source Tool Enables Local Multi-Speaker ASR & Diarization in Browser
doodlestein · x · 2026-08-13
An open-source tool named frankenwhisper is gaining attention for delivering world-class multi-speaker speech recognition, diarization, and denoising entirely locally within a web browser.
- Tech Highlights: The project rebuilds OpenAI's Whisper (809M parameters) and NVIDIA's Streaming Sortformer (123M parameters) diarizer in pure, memory-safe Rust compiled to WebAssembly. It requires no PyTorch or Python, running entirely on the CPU.
- Privacy & Security: All processing of audio and 1.3GB of model weights happens in-tab, ensuring 0 bytes are uploaded and guaranteeing privacy.
- Capabilities: It supports real-time transcription and speaker attribution for up to 4 speakers, with model weights verified by SHA-256 pins.
Related event: Franken Whisper: Open-Source Browser-Based ASR Tool(3 posts)→
More from Infra
- New hardware day: local LLM deployment the moment it arrives — lxfater · 2026-10-16
- Together AI's score centering fixes RL training collapse under quantized inference — nrehiew_ · 2026-10-16
- Inference is splitting prefill and decode across specialized chips, not just H100s — AccBalanced · 2026-10-16
- Rebasely launches in beta: Git hosting built for humans and coding agents — handotdev · 2026-10-16
- Drama3 TTS lands in audio.cpp: up to 5x real-time on RTX 5090, <8GB VRAM mode too — Acceptable-Cycle4645 · 2026-10-16
- Three CoreWeave co-founders have sold $3.3B in shares since IPO amid credit worries — SumitGup · 2026-10-16