babelscribe MCP
Local stdio MCP for offline audio and video transcription and subtitle generation using whisper.cpp, with hardware-dependent acceleration and multilingual support. First use downloads binaries and model weights; the user's media stays local during transcription, but downloaded models and fine-tunes have their own licensing. Repository created October 6, 2026.
MCP unverified
Integration
| Transport | stdio |
| Auth | none |
| Endpoint | uvx babelscribe mcp |
| Install | Register uvx babelscribe mcp as a local stdio server (Python 3.9+); allow time and disk for initial model/backend downloads |
Use Cases
| 01 | Transcribe a local interview recording |
| 02 | Produce SRT or VTT subtitles |
| 03 | Review speech text without uploading the media |
Tags
transcription subtitles whisper-cpp offline multilingual
Machine-readable: /api/servers.json
· JSON-LD schema embedded in <head>