Ffvoice
by chicogongcommunityv0.8.3
3
ffvoice-engine is an offline audio processing and speech-to-text MCP server that combines microphone/system audio capture, audio cleanup, Whisper-based transcription, live captions, and optional speaker diarization. It is designed to run locally without sending audio to a cloud API and includes CLI and Python bindings in addition to the MCP server.
From the registry: Offline speech-to-text & speaker diarization MCP server: transcribe audio on-device, no cloud
Install & configure
Prerequisites
- requiredInstall ffvoice MCP packageInstall the Python package with the MCP extra so the ffvoice-mcp command is available. Optionally include the diarization extra to enable speaker diarization tools.
https://pypi.org/project/ffvoice/$ pip install 'ffvoice[mcp]' - optionalOptional: install diarization supportInstall the diarization extra if you want the transcribe_file_with_diarization MCP tool.
https://pypi.org/project/ffvoice/$ pip install 'ffvoice[mcp,diarization]'
Install in your client
Paste into the AI you're chatting with — it will figure out the rest
Please install the `ffvoice` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install ffvoice MCP package** — Install the Python package with the MCP extra so the ffvoice-mcp command is available. Optionally include the diarization extra to enable speaker diarization tools. Run: `pip install 'ffvoice[mcp]'` (https://pypi.org/project/ffvoice/)
Optional prerequisites:
- Optional: install diarization support — Install the diarization extra if you want the transcribe_file_with_diarization MCP tool. Run: `pip install 'ffvoice[mcp,diarization]'` (https://pypi.org/project/ffvoice/)
Canonical MCP server config (stdio transport):
- command: `ffvoice-mcp`
- args: []
Note: The README provides a Claude Desktop example showing ffvoice-mcp with no args or environment variables. This is the simplest portable path. Models are downloaded automatically on first use. The server is fully offline and does not require API keys.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.Note: The README provides a Claude Desktop example showing ffvoice-mcp with no args or environment variables. This is the simplest portable path. Models are downloaded automatically on first use. The server is fully offline and does not require API keys.
Tools (5)
record
Records audio from a selected device to WAV or FLAC, with optional processing such as normalization, high-pass filtering, and RNNoise noise reduction.
transcribe
Transcribes an audio file using Whisper and can output plain text, SRT, VTT, or JSON with timestamps.
transcribe-live
Performs live transcription while recording audio, producing realtime transcription events.
live-captions
Streams partial and final caption updates during recording for realtime subtitles.
diarize
Adds speaker diarization to transcriptions so output can identify who spoke when.
Prerequisites
- • Python 3.10
- • python3
Packages
ffvoice
pypiv0.8.3transport: stdio
Details
- Language
- C++
- License
- MIT
- Released
- May 19, 2026
- Last commit
- May 19, 2026
- Registry name
- io.github.chicogong/ffvoice
Related Servers
Claude Flow
AI orchestration with hive-mind swarms, neural networks, and 87 MCP tools for enterprise dev.
68.0k
Codebase Memory Mcp
Persistent codebase knowledge graph. Survives session restarts and context compaction.
39.2k
Serena
A powerful toolkit for coding, providing semantic retrieval and editing capabilities.
28.1k