# Train in Silence

A task-aware MCP server that calculates the VRAM and compute a fine-tuning job needs and returns the cheapest and fastest matching GPUs across a dozen cloud providers, instead of comparing prices by hand.

Train in Silence is a task-aware MCP server for planning where to run an LLM fine-tuning job. Instead of just listing GPU prices, it reads a description of the training job, works out the VRAM and compute it needs, and returns configuration options across roughly a dozen cloud GPU providers ranked as cheapest, fastest or most balanced.

It can be called from Claude Code or Claude Desktop as an MCP tool, from a command-line interface with a YAML request file, or from a REST API. Provider API keys are optional: without them, it falls back to public GPU-price aggregators or bundled sample data to still produce an estimate.

It installs from PyPI and is registered with Claude Code through the standard MCP command.

Website: https://github.com/hlpun/Train-in-Silence