Features
Compress noisy command output at the source, so agents spend context on signal instead of terminal spam.
How Boost saves tokens
Boost trims repetitive chatter from stdout while keeping failures, summaries, counts, and next steps in what the model sees.
Boost keeps the command useful for humans and agents by preserving the signal: failures, summaries, counts, and next clues. The repetitive log stream gets compressed before it spends context.
Noisy test runs
Long test suites often print hundreds of passing cases and repeated stack frames. Boost keeps the failures, flaky counts, timings, and summary so the agent can act without reading the whole scrollback.
$ boost pytest -q
$ boost go test ./...
$ boost npm test
Build and lint logs
Type checkers, linters, and bundlers can flood the terminal with progress lines and duplicate diagnostics. Boost groups the useful errors by file and keeps the command result compact.
$ boost npm run build
$ boost tsc --noEmit
$ boost ruff check .
Diffs, logs, and debug output
When an agent asks for context, raw diffs and logs can dominate the conversation. Boost deduplicates noisy lines, compacts large outputs, and keeps enough structure for the agent to decide what to inspect next.
$ boost git diff
$ boost docker logs app
$ boost kubectl logs deploy/api
Token saving
Boost tracks raw versus filtered output size and estimates saved tokens per command. Run boost report to see aggregate token savings for your recent work.
To see which commands are supported, check the Supported Commands reference.
Sustainability
Every saved token is energy you didn't burn. Boost translates raw token savings into an estimated emissions impact so the work your agents skip becomes a number you can share with your team.
AI agents stream a lot of noise. Boost trims it before it becomes model inference compute. Run boost report to see an estimated CO2e avoided alongside saved tokens and dollars.
The estimate uses a configurable kg-CO2e-per-million-saved-tokens heuristic. It's a directional sustainability signal, not audited carbon accounting, but it turns context saved into a concrete reminder that less AI exhaust reaches the atmosphere.
Sample report line
You avoided 359 g CO2e.
≈ 1.7M saved tokens at 0.21 kg CO2e / 1M.
