whisper/whisper
MCP ServerWhisper MCP Server - audio transcription with OpenAI Whisper
check: npx https://www.npmjs.com/package/whisper-mcp
Whisper audio transcription MCP server. Transcribe audio, convert formats, and generate speech using OpenAI's Whisper.
MCP Server
- Package:
mcp-server-whisper (Python — use uvx) - Transport: stdio
- Command:
uvx mcp-server-whisper
Authentication
OpenAI API key
| Variable | Required | Description |
|---|---|---|
OPENAI_API_KEY | Yes | OpenAI API key |
AUDIO_FILES_PATH | Yes | Directory for audio files |
Available Tools
| Tool | Description |
|---|---|
list_audio_files | List audio files |
transcribe_audio | Transcribe audio |
convert_audio | Convert format |
compress_audio | Compress audio |
create_audio | Text-to-speech |
chat_with_audio | Chat about audio |
Configuration Example
[[mcp.servers]]
name = "whisper"
type = "stdio"
command = "npx"
args = ["-y", "mcp-server-whisper (Python — use uvx)"]
timeout_seconds = 60
env = { OPENAI_API_KEY = "your-value", AUDIO_FILES_PATH = "your-value" }
tools = []Notes: Python-based — use uvx. 11 tools total.