A minimal Go CLI tool that caches JSON API responses to avoid hitting rate limits and speed up development.

Fully cooked by Claude.

go run main.go <port> <target-url>

Examples:

go run main.go 8080 https://api.github.com
go run main.go 3000 https://jsonplaceholder.typicode.com

Then make requests to http://localhost:<port> instead of the API directly.

• Only caches GET and HEAD requests (idempotent methods)
• Stores responses as readable JSON files in ./cache/
• Handles gzip compression properly (decompresses for storage, re-compresses when serving)
• Cache key = MD5 hash of method + path + query params

• First request: Cache MISS → fetches from API and caches
• Subsequent requests: Cache HIT → serves from disk (with gzip if client supports it)
• POST/PUT/DELETE: No cache → always proxies through

The proxy perfectly mirrors the original API's compression behavior:

• If the API sends gzipped responses → cached responses will also be gzipped when served
• If the API sends uncompressed responses → cached responses remain uncompressed
• Always stores readable JSON in cache files (decompressed for inspection)

This ensures your application behaves identically with both the proxy and the real API.

• Avoiding API rate limits during development
• Working offline with previously fetched data
• Speeding up repetitive API calls in scripts
• Testing your app's gzip handling without hitting the real API

• Cache files: ./cache/*.json
• Clear cache: rm -rf cache/
• Inspect responses: cat cache/<hash>.json