MCP SaluteSpeech
An MCP server for Sber's SaluteSpeech voice API: records speech from the microphone, recognizes it and speaks the agent's reply
- 9 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Records audio from the user's microphone and sends it to an external API
- Requires a SaluteSpeech token that should be kept only in environment variables
Install
- Works in
- Services
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/mcp-salutespeechDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Records audio from the user's microphone and sends it to an external API
- Requires a SaluteSpeech token that should be kept only in environment variables
This is third-party code. Review the repository files before installing.
What it does
The server gives the agent two voice tools. One records audio from the microphone until a three-second pause, sends it to the SaluteSpeech recognition API and returns the text. The other turns the agent's reply into speech through the same API and plays it through the computer's speakers, with a choice of format and voice. It runs as a local MCP server on top of the official SaluteSpeech voice API.
Who it is for. For people who want to talk to an agent by voice instead of typing, and for developers building LLM-based voice interfaces.
Good fit when
- You want to dictate a request to the agent by voice instead of typing
- You want the agent's answer read out loud, for example hands-free work
- You need a prototype Russian-language voice interface for an agent system
Not a fit when
- You need speech recognition in languages other than Russian
- There is no microphone or speaker access on the machine running the agent
Example request
Record what I say next and read your answer out loudMCP
- Transport
- stdio
- Authentication
- API key
Environment variables
SALUTE_SPEECHrequired, secret- A Basic authorization token for the SaluteSpeech API.
Limitations
Requires a SaluteSpeech token from Sber. Recording is hardcoded to a three-second pause and is not configurable in the README. Linux and Windows may need extra PortAudio and other system audio dependencies. There have been no pushes for over a year, so ongoing support is not guaranteed.
How to disable. Remove the mcp-salutespeech block from your MCP client config.
FAQ
How long does the microphone recording last?
Until roughly a three-second pause, after which the recording is sent for recognition automatically.
Where do I get the SALUTE_SPEECH token?
In the SaluteSpeech developer account on Sber's portal, it is a Basic authorization token for the API.
Discussion
No questions or reviews yet.
Related
All Sber toolsMCP server for Sber SaluteSpeech: transcribe Russian speech to text and voice text with Sber voices
Semantic search over Jivo and SaluteBot docs
MCP Docs Hub
A collection of MCP servers with FAISS indexes for semantic search over Jivo, SaluteBot and SberSalute docs: search, file reads and contents
Sber API MCP for corporate clients
Sber API MCP
Sber's official MCP for corporate clients: an agent works with ruble payment orders and account statements through the bank API
Unofficial MCP server and Python SDK for SberBank Online: balances, operations and transfers between your own accounts with confirmation
An MCP server for image generation through Sber's official Kandinsky API, with style choice and negative prompts
Sberbank business API MCP server
@theyahia/sber-mcp
An MCP server for the Sberbank Business API: accounts, balance, statements, payment orders and counterparties for an AI agent
README in short
The Russian README explains the purpose of the two tools, lists system audio dependencies for Linux, macOS and Windows, gives the JSON for connecting the server via uvx with the SALUTE_SPEECH variable, and separately shows adding it to Cursor globally or per project. MIT license.