Matryoshka
Active·★ 150·Apache-2.0·Updated 2026-05-17
Matryoshka (RLM) addresses the limitation of fixed LLM context windows by using a recursive language model approach. Instead of chunking or RAG, the LLM outputs commands in a constrained symbolic language called Nucleus, which are executed by the Lattice logic engine. This reduces entropy, enables safe execution, and achieves 97% token savings through handle-based storage.
#ai-assistant#document-analysis#llm#llm-tools#mcp#mcp-server#mcp-servers#model-context-protocol
01
Features
01Recursive Language Model (RLM) architecture based on a paper
02Nucleus DSL: constrained S-expression language for safe LLM output
03Lattice engine: parser, type inference, constraint resolver, and solver
04In-memory handle storage with SQLite FTS5 achieving 97% token savings
05Tree-sitter code-aware querying for structural symbols
02
Compatibility
Node.js
Node.js
Verified via docs
03
Quick start
1
$ npm install -g matryoshka-rlm
04
Use cases
↳Document analysis: search, filter, count, and sum large log files
↳Code analysis: extract functions, classes, find references in source code
↳Large-scale data queries without exceeding LLM context limits
05
Alternatives
chiasmus★ 217
Chiasmus is an MCP server that gives language models access to formal verification
UNO-MCP★ 14
Unified Narrative Operator MCP server — AI-powered prose expansion, repetition analysis & narrative enhancement
reddit-mcp-buddy★ 862
Clean, LLM-optimized Reddit MCP server. Browse posts, search content, analyze users. No fluff, just Reddit data.
mcp-ts-core★ 153
TypeScript template for building Model Context Protocol (MCP) servers. Ships with declarative tools/resources, pluggable auth, multi-backend storage, OpenTelemetry observability, and first-class support for both local and edge (Cloudflare Workers) runtimes.
Comments
Log in to leave a comment
No comments yet. Be the first!