Quick Install

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

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "claude-runway": {
      "command": "uvx",
      "args": [
        "claude-runway"
      ]
    }
  }
}

Or install with pip: pip install claude-runway

README Excerpt

Local-compute token savings for Claude Code. Two independent pieces, both built on the same idea: push work onto local compute (a local vector DB, a local LLM) instead of Claude's own context, so Claude only pays tokens for a distilled result rather than raw data. 1. **Qdrant codebase memory** — per-project semantic memory over code/docs, so Claude can retrieve relevant chunks instead of grepping and reading whole files, and persist distilled findings across sessions.

Tools (12)

CLAUDE_RUNWAY_CACHE_DBCLAUDE_RUNWAY_COMPRESS_THRESHOLD_CHARSCLAUDE_RUNWAY_LMSTUDIO_MODELCLAUDE_RUNWAY_LMSTUDIO_URLCLAUDE_RUNWAY_SAVINGS_DBCLAUDE_RUNWAY_TRACK_SAVINGSFASTEMBED_CACHE_PATHINDEX_EXCLUDE_DIRSINDEX_INCLUDE_EXTENSIONSModuleNotFoundErrorcommandfind_in_collection

claude-runway

Quick Install

README Excerpt

Tools (12)

Related AI/LLM Servers