Nestbolt comparisons

We make Nestbolt. Two pages compare it with other nesting tools. The Deepnest alternative page gives the material use measured in Nestbolt's frozen benchmark against Deepnest, SVGnest and Sparrow, with the date of the run. The nesting for AI agents page lists which nesting tools an AI agent can call through an MCP server, a command line or a library; it says which numbers we measured and which we read from the other tools' own pages, and when it was last checked.

A Deepnest alternative, with the benchmark to inspect

In Nestbolt's frozen v1 benchmark (build 75cbb32, final acceptance run of 2026-10-13), Nestbolt used 8.5% less material at the median than the best of Deepnest, SVGnest and Sparrow at equal 60-second budgets, and 9.6% at 300 seconds on the long subset. Sparrow, which Nestbolt builds on, still wins 8 of the 56 jobs at 60 seconds. The page explains how the benchmark was run and where Sparrow wins. The results page has every per-job number.

Nesting for AI agents: Nestbolt, OpenNest, pyFit, Sparrow and Deepnest compared

Checked 2026-10-20. Reading the tools' repositories, four MCP directories and GitHub search, we found three open-source nesting tools that an AI agent can call over MCP: Nestbolt (5 tools, SVG and DXF in), OpenNest by ajisaacs (19 tools, DXF in, built from source) and pyFit (4 tools). SmartCut is a hosted, paid MCP service. Sparrow, CLI-nesting and sheetnest have command lines, while Deepnest and SVGnest have none. We ran only Nestbolt: three fresh v0.1.0 MCP servers nested its example job six times, placing 32 of 32 parts on one sheet each time, and all six layouts were byte-identical. Everything about the other tools comes from their own documentation.

The hosted browser app runs a preview build from main; stable v0.1.0 remains a download.

Comparisons for our other tools