ABOUT SCORECOMPUTE

A place for AI assistants
to calculate.

ScoreCompute is a remote MCP server that gives AI assistants 38 tools, including 22 scientific engines written in Rust. Connect at https://scorecompute.com/mcp; public access needs no API key. Results depend on the inputs and model you supply.

What it does

An assistant calls a tool through the Model Context Protocol. A bounded Rust engine does the calculation and returns the result with its assumptions and limits. People can run the same tools from the interactive demonstrations.

How results are checked

Each scientific engine has an evidence card listing its references, tests and source hash. Studies that did not meet their acceptance criteria are published on the trust page instead of being removed. Evidence covers the listed cases, not every possible input.

What does “bounded” mean?

Every tool has a fixed maximum input size and amount of work, stated on its engine page. A request outside those limits is refused, and a long calculation ends with an explicit status or a timeout instead of running indefinitely.

Which ScoreCompute tool runs Monte Carlo on my own model?

Use propagate_uncertainty for a measurement model and simulate_plan for uncertain task durations. simulate_pi only estimates π.

What it is not

A result is conditional on the supplied inputs and model. It is not a certification, a regulatory approval or a substitute for reading the method.

Who runs it

ScoreCompute is part of the ScoreZenith ecosystem and is published by TENDIL COURTAGE. See the legal notice.

Documentation

Connection guide and recorded examples on GitHub. The engines run on the hosted service; their source is not published there.

Contact

contact@scorezenith.com · Suggest an engine · Service status

First published 2 November 2026 · Last updated 2 November 2026