Back to Browse

Meaning Model MCP Server

Developer ToolsLow Risk10.0MCP RegistryLocal
Free

Server data from the Official MCP Registry

General-purpose, revisable process models with application-defined categories. Rust engine required.

About

General-purpose, revisable process models with application-defined categories. Rust engine required.

Security Report

10.0
Low Risk10.0Low Risk

Valid MCP server (2 strong, 1 medium validity signals). No known CVEs in dependencies. Package registry verified. Imported from the Official MCP Registry. Trust signals: trusted author (3/3 approved).

6 files analyzed · 1 issue found

Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.

What You'll Need

Set these up before or after installing:

Absolute path to the matching life-sim-engine executable. Install this npm version and run meaning-model-mcp --install-engine for a supported release, or --build-engine with Cargo; see the installation guide.Optional

Environment variable: LIFE_SIM_ENGINE_BIN

How to Install

Add this to your MCP configuration file:

{
  "mcpServers": {
    "io-github-emergent-wisdom-meaning-model": {
      "env": {
        "LIFE_SIM_ENGINE_BIN": "your-life-sim-engine-bin-here"
      },
      "args": [
        "-y",
        "@emergent-wisdom/meaning-model-mcp",
        "--yes"
      ],
      "command": "npx"
    }
  }
}

Documentation

View on GitHub

From the project's GitHub README.

The Meaning Model

Construct worlds coarsely, then develop the parts that matter. The Meaning Model connects world records, concepts, descriptions, understanding, and story text in one addressed representation. Finer detail must respect an accepted account or revise it explicitly.

This repository brings together the theory, its compact grammar appendix, the Rust engine and MCP interface, and The Book of Conditions, a worked construction with a complete twelve-chapter manuscript.

See the changelog for release changes, upgrade notes, and unreleased work.

Choose the model your application needs

Model only what serves your purpose. Choose the processes, categories, relationships, and level of detail; use a supplied template, adapt its output, or author a model directly. A single process is a valid starting point. A browser, learning tool, or relationship application need not build the same account of a person, or a complete person model at all.

The intended use is an external model an AI can work with: connect observations across time, compare explanations, explore possible developments, and decide what to investigate next. The modeler supplies those interpretations; the tool preserves their structure and revisions and executes supported, declared laws. Categories and decompositions are part of what the modeler develops and tests, not a fixed inventory to fill in. Shared record and validation rules remain in force as the application vocabulary changes.

See application-specific categories for template choices and a runnable vocabulary-revision example.

Read first

The paper and grammar appendix were revised on September 12, 2026. The Zenodo series provides the archived versions.

The six world-record forms are Concept, Thing, Event, Binding, Cut, and Realization. A Cut divides one declared unit among exclusive sibling answers and an explicit remainder; its weights sum to one. Ordinary process links are unweighted. Dates, money, counts, and physical measurements keep their external units. Understanding Nodes and Document Nodes share the address space without acquiring the authority of accepted world facts.

Try it

The smallest demonstration requires only Node.js:

node --test examples/refinement-trial/example.test.mjs

It shows a refinement that fits its parent, a locally valid refinement that contradicts the parent, and an explicit revision. The example is synthetic, not evidence that the full implementation or learning proposal is complete.

For the Rust engine and MCP tools, install Rust/Cargo, Node.js 22.18 or newer, and npm, then run:

make install
make build
make test

For a native example, run node examples/progressive-authoring/run.mjs after building. Progressive authoring in Rust opens a Cut after accepted history, rejects incompatible detail, completes a partial temporal contract, and explicitly revises and continues the same world.

Start the MCP server after building:

cd mcp-server
npm start

See the MCP guide, engine guide, and modeling protocol. Existing life-sim command, schema, resource, and binary identifiers are retained for compatibility with saved artifacts. They do not identify a second engine.

The package name is @emergent-wisdom/meaning-model-mcp. Its official MCP Registry identity is io.github.emergent-wisdom/meaning-model, described by server.json. Registry-based clients must supply the path to an explicitly installed or built engine. Install the engine matching the package version. meaning-model-mcp --install-engine explicitly downloads and verifies that version's published release engine. Alternatively, use meaning-model-mcp --build-engine to build the included Rust source. Neither npm installation nor normal server startup downloads or builds an engine. To prepare a local npm tarball, run make npm-package. See the package installation guide. Creating a tarball does not publish it or establish ownership of the npm scope.

What is demonstrated

The Book is a bounded example of model-assisted authorship: macro accounts, complete lives, local Cuts, event descriptions, linked construction rationales, and prose were developed and revised together. Its accepted sources can be imported into Rust and the manuscript rendered from document nodes. That retrospective import is not a recovered transaction log of the original authoring process.

The engine implements typed records, immutable revisions, persistence, candidate acceptance, normalized Cuts, and bounded narrative graph operations. Optional temporal Cut contracts check explicit answer projections and duration mixtures, including whether partial detail leaves a feasible remainder. An accepted world can move to its direct next model revision after time has advanced, with an explicit refinement or revision, a compare-and-swap check, and an immutable receipt preserving both heads. Portable narrative training and project checkpoint exports across that revision boundary remain unsupported; the session database retains the complete history. It does not implement every contract in the paper. The implementation boundary distinguishes implemented checks, authoring conventions, and missing behavior. The preregistered matched comparison and independent read-back study remain prospective. Neither a passing validator nor an authored numerical portrait establishes historical truth, psychological validity, or an advantage over ordinary writing.

Repository map

PathContents
paper/Paper and appendix sources, shared style, figures, bibliography
rust-engine/Shared Rust implementation, optional numerical simulation capabilities, tests and examples
mcp-server/MCP interface to that engine
profiles/Reusable authoring conventions, not mandatory human categories
docs/Operational documentation and implementation limits
examples/refinement-trial/Small standalone construction example
examples/progressive-authoring/Native temporal refinement, explicit revision, and continued world history
examples/book-of-conditions/Accepted Book, model sources, rationale nodes, and reproducible export
output/pdf/Ready-to-read paper, grammar appendix, and Book PDFs

Meaning Model owns the representation and joint construction method. The separate Life Simulation project studies process-history generation, inference, learning, and their empirical evaluation. It consumes this engine; it does not maintain a second implementation. A digest-bound copy of its paper is included only as a companion MCP reading resource under docs/companions/life-simulation/.

Build the documents

With latexmk and a LaTeX installation containing the imported packages:

make paper
make grammar
make book

To check the code, examples, resources, and documents together:

make check

Release packaging

make release checks and builds the package, then exports an allowlisted clean directory beneath build/. The export includes the three PDFs and their sources, but not .git, build caches, private planning, review transcripts, or earlier constructions.

See the release guide for packaging and provenance details. Exporting files does not push to GitHub or publish a package.

License and citation

Original code is MIT licensed. Original papers, documentation, authored model data, and the Book are CC BY 4.0. NOTICE preserves the boundary for third-party material. CITATION.cff identifies the author and preferred paper citation; it includes the repository URL and the Meaning Model's Zenodo concept DOI. The DOI identifies the paper's version series; the release manifest identifies the exact files.

Reviews

No reviews yet

Be the first to review this server!