Homer Studio
A local-first voice and audiobook studio that brings neural narration, voice conversion and precise audio editing into a single desktop workflow.
The whole journey from manuscript to audio
Homer Studio connects chapter import, narration, review and export. Authors can work with source text, voice references and recorded takes without moving between unrelated generation tools and audio editors.
AI belongs inside an editable workflow
Local Chatterbox workers support speech generation and voice conversion. Generated takes remain selectable and reversible. A waveform editor provides split, trim, voice assignment and mixing controls, with video references and soundtrack placement for audiovisual work.
Native execution, portable projects
A React and TypeScript interface sits on shared Rust services, with Tauri for macOS and an authenticated self-hosted browser edition on Linux. FFmpeg handles rendering. Portable .wavehs projects retain source assets; models are installed explicitly and execution stays local or on the user-managed server.
Engineering beyond the first successful generation
Persistent projects, cancellable work, waveform caching, export validation and worker cleanup make the difference between a model demo and a practical editing application. WAV, M4A and MP4 exports connect the studio to the rest of a production pipeline.