shared-context-cache-mcp-server

MCP server for shared context caching — AI agents share computed results to reduce token cost and latency.

Why?

Every AI agent constantly re-computes the same results: weather lookups, price checks, document summaries, research queries. With this MCP server, agents share their computed results through a common cache.

Network effect: More agents caching → more cache hits → everyone saves tokens and latency.

Like a CDN, but for agent intelligence.

Install

pip install shared-context-cache-mcp-server

Tools

Tool	Description
cache_lookup	Look up a cached result by key — check before computing
cache_search	Search cache by keywords — find relevant precomputed results
cache_store	Store a computed result for other agents to reuse
cache_stats	Get cache performance stats (hits, misses, cost savings)
cache_list	List available cache entries, optionally filtered by tags

Usage Pattern

1. Before computing: cache_search("weather berlin") → cache_lookup("weather:berlin:today")
2. Cache hit? → Use the stored result directly, no API call needed
3. Cache miss? → Compute the result, then: cache_store(key, value, tags="weather,berlin")
4. Other agents now benefit from your computation

Claude Desktop Config

{
  "mcpServers": {
    "shared-context-cache": {
      "command": "shared-context-cache-mcp-server"
    }
  }
}

Cache Key Conventions

Use descriptive, hierarchical keys:

• weather:berlin:2026-03-28
• research:arxiv:2501.00001:summary
• price:bitcoin:usd:2026-03-28
• analysis:company:AAPL:q1-2026

Backend

Powered by agent-apis.vercel.app/api/cache — a shared cache API built for the agent economy.

License

MIT — AiAgentKarl