l0-cache

Appearance

l0-cacheCLI proxy for LLM token savings

Reduce AI coding assistant token consumption by 50-80% with a single command prefix.

Get Started
Architecture
l0-cache --stats telemetry dashboard

Universal Filtering

Works with any command. No per-tool parsers and no alias rewriting -- just prefix with l0-cache, or enable the optional Claude Code hook.

Zero Overhead

Single-threaded, synchronous Rust binary. Sub-millisecond overhead. 700 KB on disk.

Production Hardened

UTF-8 lossy reads, OOM protection, SIGPIPE handling, POSIX exit codes, metrics rotation.

Cross-Platform

macOS, Ubuntu, Alpine, LXC, Proxmox, VPS. Static musl build for containers.