Context Proxy

MCP proxy reducing tool schema token overhead by 6-7x through lazy-loading, caching, and response compression.

GitHub
Category MCP Servers
Language TypeScript
License NOASSERTION
Added Mar 28, 2026
Views 38

About

Sits between MCP clients and servers to optimize token usage. Returns compressed stubs initially, lazy-loads full schemas only when tools are invoked, caches schemas to disk with 24h TTL, deduplicates identical schemas, and compresses large responses. Includes verifiable savings proof logging and auditable metrics. Supports HTTP/SSE transport.

Is this your product? Claim this listing with an email at your domain to post updates.

Is this your project?

Claim this listing to manage your page, access analytics, and unlock upgrades. Verification takes 60 seconds.

Log In to Claim

Share This Project

Share on X Share on LinkedIn

Embed Badge

Add this badge to your README:

[![Listed on AiList](https://hifriendbot.com/ai-list/badge/context-proxy.svg)](https://hifriendbot.com/ai-list/context-proxy/)
Listed on AiList

Compare

Context Proxy vs mwilliams_mcpbridge Context Proxy vs MCP Lazy Load Context Proxy vs Toroidal Toolshed

Similar Projects

mwilliams_mcpbridge

3

MCP Bridge - 99% context reduction for AI agents. Consolidates multiple MCP servers behind lazy-loading meta-tools. Result compaction, schema caching, health checks.

MCP Servers

MCP Lazy Load

10

Proxy that lazy-loads MCP tools on demand to reduce context window token usage by 90%.

MCP Servers TypeScript

Toroidal Toolshed

4

Proxy that multiplexes multiple backends through two tools, reducing token overhead.

MCP Servers Python

List Your Project

Join the directory Ai agents read. Free forever.

Submit Your Project