Caveman.Mcp 1.4.4

Caveman.Mcp

MCP (Model Context Protocol) server for Caveman — exposes NLP token compression, language detection, content routing and summarization as MCP tools for Claude Code, OpenCode, Cursor, Windsurf, and any MCP-compatible agent.

But for Claude and other Agents we recommend our dedicated Synthelion project which can be found here https://github.com/francescopaolopassaro/Synthelion. This MCP was created to have a package to remain in the C#/Caveman environment for future developments facilitated for this kind of purpose. So if you want to use Caveman as an independent MCP and plugin for your agent, use Synthelion.

Part of the Caveman ecosystem: NLP prompt compressor for LLMs. Up to 70% token reduction, 50+ languages, zero ML models.

Technology Partnership

<img src="https://www.digitalsolutions.it/img/partners/novaroutelogo.png" alt="NovaRouteAI" height="180" style="max-width: 100%; height: auto; min-height: 180px; max-height: 190px;">

NovaRouteAI — Build with Chinese AI models through one simple API.

NovaRouteAI helps developers and AI SaaS teams test, compare, and run models like DeepSeek, Qwen, Doubao, Kimi, and GLM without managing multiple provider accounts. Start with test credits and optimize your cost per successful task.

Click here to know NovaRouteAI

Quick install — Claude Code

{
  "mcpServers": {
    "caveman": {
      "command": "caveman-mcp"
    }
  }
}

Or with dotnet tool:

dotnet tool install --global Caveman.Mcp

Tools

Usage

compress

compress(text: "I would like to know if it is possible to receive information.", level: "semantic")
→ { "compressed": "know possible receive information", "efficiency_pct": 65.0, "energy_mwh": 0.05, "co2_mg": 0.02 }

route_content

route_content(content: "[{\"name\":\"Alice\",\"age\":30}]", profile: "balanced")
→ { "compressed": "| name | age |\n| Alice | 30 |", "strategy": "JsonCrush:MarkdownTable", "savings_pct": 47.0 }

retrieve

retrieve(documents: ["Tesla ships a new battery", "Rivian raises truck price", "GDP grew 2%"], query: "car", topK: 2)
→ [{ "index": 0, "document": "Tesla ships a new battery", "score": 1.83 }, ...]

skeletonize_code

skeletonize_code(code: "public int Add(int a, int b) {\n    // sums two numbers\n    return a + b;\n}")
→ { "compressed": "public int Add(int a, int b) {\n    /* ... */\n}", "functions_skeletonized": 1, ... }

near_duplicate

near_duplicate(a: "User bob logged in from 10.0.0.1", b: "User alice logged in from 10.0.0.2")
→ { "is_near_duplicate": true, "hamming_distance": 2, "max_distance": 3 }

compress_sql

compress_sql(sql: "SELECT  id, name\nFROM     users\nWHERE    age > 18;")
→ { "compressed": "SELECT id, name FROM users WHERE age > 18;", "was_compressed": true, "tuples_dropped": 0, "ccr_hash": null }

compress_sql(sql: "INSERT INTO t VALUES (1,'a'),(2,'b'),(3,'c');", foldValues: true)
→ { "compressed": "INSERT INTO t VALUES (1,'a'),(2,'b'),(3,'c'); -- CCR:3f2a9b1c8d4e", "was_compressed": true, "tuples_dropped": 0, "ccr_hash": "3f2a9b1c8d4e" }

idf_languages

idf_languages(iso3: "eng")
→ { "language_count": 56, "languages": [ "ben", "bul", "cat", ... ], "iso3": "eng", "has_data": true, "corpus_size": 1150000000 }

Supported languages (50+)

Afrikaans · Arabic · Armenian · Basque · Bengali · Bulgarian · Catalan · Chinese · Croatian · Czech · Danish · Dutch · English · Estonian · Finnish · French · Galician · German · Greek · Hebrew · Hindi · Hungarian · Icelandic · Indonesian · Irish · Italian · Japanese · Kannada · Kazakh · Korean · Latin · Latvian · Lithuanian · Macedonian · Malay · Marathi · Norwegian · Persian · Polish · Portuguese · Romanian · Russian · Serbian · Slovak · Slovenian · Spanish · Swedish · Tamil · Telugu · Thai · Turkish · Ukrainian · Urdu · Vietnamese

Links

• Caveman core: https://github.com/francescopaolopassaro/caveman
• NuGet (core): https://www.nuget.org/packages/Caveman
• Python port (Synthelion): https://pypi.org/project/synthelion/

© 2026 Passaro Francesco Paolo — Digitalsolutions.it