Vibe Coding Discovertoken-optimizer-mcp
View on GitHubMeasure token savings per AI coding agent, optimize context, and share a live local knowledge graph across 16 CLI clients.
MCP server and plugin that compresses AI coding agent context with cache-aware token optimization, enforces diff-only re-reads, and keeps a live per-project knowledge graph. Adds per-agent token accounting, a local savings dashboard, and benchmarks. Install via Claude Code plugin or npx; no telemetry.
Use Cases
Cut AI coding agent token spend via cache-weighted context compressionEnforce diff-only re-reads instead of full file dumpsPersist per-project findings in a local knowledge graph and re-inject them on file touchAttribute token savings per agent (Codex, Claude Code, Gemini)Expose a live local dashboard of verified net token savingsMeasure MCP progressive-disclosure savings before/after expansionIntegrate the same optimizer across 16 CLI clients via wrapper and hooksRank the biggest per-session token costs into a single audit queue
Built With
- Language
- JavaScript
- Frameworks
- Node.js · MCP SDK · Claude Code plugins · TypeScript · Jest
Tags
mcp · mcp-server · token-optimization · context-compression · prompt-caching · knowledge-graph · cost-tracking · claude-code · gemini-cli · codex · local-first · hooks · benchmark · observability · cache-aware · cli-integration