Distill trims noisy CLI output for LLMs
distill pipes command output through a local summarizer so agents get only the signal they need instead of walls of logs, diffs, and stack traces. The open-source CLI supports local and OpenAI-compatible providers, making it a practical fit for Codex, Claude Code, and similar workflows.
Distill addresses a real agent bottleneck: LLMs are much better at reasoning over compact facts than parsing walls of logs, diffs, and stack traces. Its local and OpenAI-compatible providers make it easy to keep summarization close to the machine, though exact-output workflows still need a raw-output bypass when fidelity matters.
DISCOVERED
67d ago
2026-03-21
PUBLISHED
67d ago
2026-03-21
RELEVANCE
AUTHOR
Github Awesome