token-saver

View on GitHub

Content-aware output compression for AI coding assistants. 36 specialized processors cut CLI output tokens by 60-99% (git, pytest, npm, terraform, kubectl, docker, and more) without losing errors, diffs, or stack traces.

★ 15213 forksPythonApache-2.0ppgranger

Local, deterministic CLI-output compressor that sits between your terminal and AI coding assistants, using 36 format-aware processors to cut tool output tokens by 60-99% while preserving errors, diffs and stack traces. Installs as a Claude Code / Antigravity plugin; no network calls or extra model inference.

Use Cases

Reduce input tokens in Claude Code sessionsCompress pytest/jest test output before it reaches the modelTrim git diff and git log output to keep changes onlyCollapse npm/pip/cargo install and build logsShrink terraform plan, kubectl, docker outputRedact secrets in env/printenv outputTrack token savings across agent runsDetect new vs repeated test failures between runs (Delta)Run in CI pipelines feeding CLI output to an LLM

Built With

Language: Python
Frameworks: Claude Code · Antigravity CLI

token-saver

Use Cases

Built With

Tags