# debpalash/voicestudio

> VoiceStudio is an open-source, fully local AI voice studio for voice cloning, voice design, video dubbing, dictation, transcription and audiobook creation in 646 languages. It takes a clean reference recording and a script, then runs them through cloning, text-to-speech and speech-recognition engines to produce speech in the cloned voice. The project is written mainly in Python, with an Electron desktop shell.

Source: https://github.com/debpalash/voicestudio
Indexed: commit 4d4cb0b on 2026-10-30 by Repowise. Facts come from the source and git history, not from the maintainer.

## Facts

- Type: MCP server
- Language: Python
- Stars: 116,236 at index time
- License: AGPL-3.0
- Code health: 7.2/10, Good
- Last commit: 2026-10-29
- Contributors: 120

## Install

From the README: https://github.com/debpalash/voicestudio#get-started

Clone the repository from GitHub.
    git clone https://github.com/debpalash/VoiceStudio.git

Change into the project directory.
    cd VoiceStudio

Install the dependencies and launch the app with Electron.
    bun run dev

## How it works

Requests arrive over HTTP and WebSocket at the backend directory, which turns them into transcripts, speech and stored project data. Each text-to-speech and speech-recognition engine runs in its own process under backend/engines, so the main application sends text or audio in and gets synthesized audio or a transcript back. backend/core provides shared services that decide where data and settings live, keep secrets out of diagnostic reports and turn failures into messages a user can act on. The electron directory is the desktop shell, which sends requests to the local backend and stays usable when the backend is slow or unreachable. The omnivoice directory turns written text into speech, can copy the voice of a short reference clip, and holds the command-line tools and web demo. The scripts directory contains the command-line tools that maintainers run around releases, store listings and local setup.

- .dagger/src/voicestudio/__init__.py: entry point
- .dagger/src/voicestudio/main.py: entry point
- backend/engines/_asr_sidecar/main.py: entry point
- backend/engines/_echo/main.py: entry point
- backend/engines/audiocpp/bootstrap.py: entry point
Modules: electron, tests, backend, mobile, scripts, omnivoice, native, .github. 4,347 files, 558,430 lines, 36,385 symbols.

## Code health

- Score: 7.2/10 (Good)
- Rank: #131 of 4,000+ indexed repos
- Hotspots: backend/services/model_manager.py, backend/services/tts_backend.py, backend/api/routers/dub_core.py, backend/api/routers/generation.py, backend/services/asr_backend.py
- Lowest file: backend/services/asr_backend.py
- Dead code: none found

## Maintenance

- Last commit 2026-10-29, first commit 2026-07-11, 3,645 commits
- 120 contributors, bus factor 1

## Security

- Security findings are shown to signed-in users at https://repowise.dev/repo/debpalash/voicestudio/security

## More

- Page: https://repowise.dev/repo/debpalash/voicestudio
- Ask this codebase: https://repowise.dev/repo/debpalash/voicestudio/chat
- Docs: https://repowise.dev/repo/debpalash/voicestudio/docs
- MCP (signs in with Repowise on first use):
    claude mcp add --transport http repowise https://api.repowise.dev/mcp/debpalash/voicestudio
