# Nomograph Labs

> We hunt for leverage in complex systems: the rare points where a small change moves the whole. We build open tools, measure with reproducible benchmarks, and keep what holds. The consistent finding across our work is that how you present information to a model matters more than how you retrieve it.

## Pages

- [Projects](https://nomograph.ai/projects): The estate on GitLab. Tools, the SysML v2 toolchain, and benchmarks. All MIT-licensed, each linking to its repository.
- [About](https://nomograph.ai/about): What the lab focuses on (context engineering, spec management, reproducible evaluation), how it started, the papers in progress, and the people.

## Projects

- [synthesist](https://gitlab.com/nomograph/synthesist): Specification graph manager for AI-augmented projects. Rust.
- [muxr](https://gitlab.com/nomograph/muxr): Harness session multiplexer for AI coding workflows. Rust.
- [jig](https://gitlab.com/nomograph/jig): Agent-shape testing harness; measures how an agent uses a tool CLI, scored by an LLM judge. Rust.
- [extension](https://gitlab.com/nomograph/extension): WASM (component-model / WIT) extension substrate for the nomograph tools.
- [pipeline](https://gitlab.com/nomograph/pipeline): CI/CD component catalog for all nomograph builds.
- [sysml](https://gitlab.com/nomograph/sysml): SysML v2 CLI and MCP server. Query, validate, and render engineering models. Rust.
- [tree-sitter-sysml](https://gitlab.com/nomograph/tree-sitter-sysml): Tree-sitter grammar for SysML v2, with bindings for Rust, Node, Python, Go, Swift, and C.
- [kebnf](https://gitlab.com/nomograph/kebnf): Convert OMG KeBNF grammars to ANTLR4 and tree-sitter. Rust.
- [sysml-bench](https://gitlab.com/nomograph/sysml-bench): Evaluation harness for AI on SysML v2 modeling tasks. Python.
- [sysml-bench-results](https://gitlab.com/nomograph/sysml-bench-results): Raw benchmark results and statistical analysis.

## Papers

Papers are in progress; each repo opens with its release. Representation Over Retrieval, on how presentation beats retrieval for AI on structured engineering tasks, was submitted to GVSETS 2026.

## Reference

- [reference.nomograph.ai](https://reference.nomograph.ai): Principles, design language, templates, and enforceable standards for the estate. The source of truth this site consumes its design tokens from.