ghx

ghx, published by Bruno Borges, is a caching proxy for the GitHub CLI (gh) designed to eliminate redundant API calls, prevent rate limiting, and dramatically speed up repeated commands. The software is currently at version 1.5.5, with five versions released in total. Positioned in the developer tools and API caching category, ghx is built primarily for AI agent workflows in which multiple agents — such as Copilot CLI, coding agents, and MCP servers — issue the same gh commands simultaneously, but it is equally useful for any environment where GitHub CLI commands are executed repeatedly. The tool operates as a drop-in replacement for the GitHub CLI: users simply invoke ghx instead of gh, and the proxy handles caching transparently. According to the publisher, cached responses are roughly ten times faster, returning in approximately 0.1 seconds compared to about 1 second for uncached calls. A central feature is singleflight coalescing, which ensures that when multiple agents request the same data at the same time — for example, five or ten agents issuing an identical command — only a single API call is actually made to GitHub, with the result shared among all requesters. This coalescing behavior directly reduces API quota consumption and helps avoid rate limiting under concurrent workloads. Safety is addressed through an allowlist-based design, meaning ghx only caches known-safe, read-only commands. When mutating operations occur, an auto-invalidation mechanism flushes related cache entries so that subsequent reads do not return stale data. For observability, ghx includes a web dashboard that displays real-time cache hit rates, per-command statistics, and a request log, giving users insight into how effectively the cache is performing. Installation is also streamlined: ghx does not require the GitHub CLI to be present beforehand, as it automatically downloads gh on first use if it is not already installed on the system.

Tags: