Train-in-Silence

Overview

Train in Silence is a specialized MCP server designed to streamline the process of selecting optimal GPU configurations for LLM fine-tuning. It analyzes training requirements and automatically identifies the cheapest, fastest GPU options across multiple cloud providers. The server implements a sophisticated architecture that transforms YAML requests into hardware recommendations through an estimator, market aggregator, optimizer, and ranking system. It supports multiple access methods including CLI, REST API, and integration with Claude Code/Desktop.

Try asking AI

After installing, here are 6 things you can ask your AI assistant:

• recommend

  Find the best GPU options for fine-tuning an LLM based on requirements

  Try:Find me the best GPU options across Vast.ai, RunPod, and Lambda for my fine-tuning jobTry:I want to run the fine-tune code in my current directory, and finish it within 20 hours

Note: Tool name inferred from CLI command 'tis recommend' and usage examples

Comparable tools

Installation

pip install train-in-silence
claude mcp add tis --scope user -- tis-mcp

pip install train-in-silence

uvicorn tis.api.server:app

{
  "mcpServers": {
    "tis": {
      "command": "python",
      "args": ["-m", "tis.mcp"]
    }
  }
}

FAQ

What cloud providers does TIS support?

TIS aggregates pricing from over 10 cloud providers including Vast.ai, RunPod, AWS, CoreWeave, Lambda Labs, Tensordock, Vultr, GCP, Azure, OCI, Nebius, CloudRift, Cudo Compute, and Verda.

Is API authentication required?

API keys are optional. If not provided, TIS automatically falls back to universal live aggregators (GPUHunt/GPUFinder) or bundled sample data.

How accurate are the GPU recommendations?

Recommendations are based on live market data when available and clearly indicate the source of truth (live:official, live:gpuhunt, live:gpufinder, or sample). The estimation model is currently in development with planned calibration improvements.

Compare Train-in-Silence with