Improvements
- Codex conversations using OpenAI-compatible Chat Completions relays now keep their prompt cache between turns, reducing token usage and keeping requests on warm caches.
- Cache hit information is now read from OpenAI, DeepSeek, Moonshot, OpenRouter, and Claude-style relay responses, making usage and cache-aware routing more accurate.
- If a provider rejects the cache key as an unsupported field, the gateway automatically retries without it and remembers not to send it to that provider again.
Bug Fixes
- Fixed cache hits showing as zero for DeepSeek-style relays, which previously prevented affinity routing from recognizing warm caches.
Install
Download from usemagpie.ai, or curl -fsSL https://usemagpie.ai/install.sh | sh. An installed magpie updates itself.
- macOS (Apple Silicon):
magpie-darwin-arm64.dmg - macOS (Intel):
magpie-darwin-amd64.dmg - Windows:
magpie-windows-amd64.exe(ARM:magpie-windows-arm64.exe) - Linux:
magpie-linux-amd64/magpie-linux-arm64(needs GTK 3 and WebKitGTK 4.1) - Terminal only:
magpie-cli-<os>-<arch>