AudioGhost AI: Extract Specific Audio Sources Using Natural Language Prompts
tom_doerr · x · 2026-08-13
A new open-source project on GitHub called AudioGhost AI allows users to extract or remove specific audio sources using natural language prompts like 'vocals', 'drums', or 'a dog barking'.
Powered by Meta's SAM-Audio model, the project features:
- Memory Optimization: Lite mode reduces VRAM usage from 11GB to 4GB
- Video Support: Upload videos to extract/remove audio sources
- Modern UI: Glassmorphism design with waveform visualization and real-time progress tracking
- Stem Mixer: Preview and compare original, extracted, and residual audio
More from coding & agent
- Qwen3.8-2.4T-A95B for long-running agents: 256K context, self-testing, function calling — togethercompute · 2026-08-13
- Qwen3-Coder 30B runs locally in GitHub Copilot at ~85 tokens/s on 96GB VRAM — ollama · 2026-08-13
- Claude Code CLI Update: Introduces Plugin Marketplace and Streaming Fixes — ClaudeCodeLog · 2026-08-13
- Beyond Trivialities: How LLMs Enable Deeper CS Learning — generativist · 2026-08-13
- AI Engineer World's Fair Highlights: Memory and Continual Learning for Agents — Stefania_druga · 2026-08-13
- The AI Agent email dilemma: How to handle replies from external clients — JanJanJaJa · 2026-08-13