turboquant-compressor

kosoymiki/turboquant-compressor
★ 5 stars C 🤖 AI/LLM Updated 1d ago
The world's first release for Ardeno. Our own custom Mesa Fork. TurboQuant KV Cache Compression - FWHT Rotation + Lloyd-Max Beta Codebook + Residual QJL. MCP Server with 14 instruments.
View on GitHub →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "turboquant-compresso": {
      "command": "npx",
      "args": [
        "-y",
        "kosoymiki/turboquant-compressor"
      ]
    }
  }
}

README Excerpt

**Termux-first MCP server for compressed local vector search, context-pack retrieval, KV/cache analysis, and Adreno/OpenCL readiness forensics.** TurboQuant Compressor is a local **Model Context Protocol (MCP)** server designed for Claude Code / stdio-style agent hosts on Android Termux and Linux. It provides compressed vector storage/search, context-pack construction, cache planning, prompt-cache linting, KV analysis, backend probing, and OpenCL/Adreno diagnostics.

Tools (13)

turboquant_adreno_loader_probeturboquant_backend_probeturboquant_cache_planturboquant_cli_mcp_profileturboquant_compressturboquant_context_pack_buildturboquant_context_pack_searchturboquant_cost_analyzeturboquant_kv_analyzeturboquant_opencl_probeturboquant_prompt_cache_lintturboquant_quantizeturboquant_vector_search