quor 0.5.0

A rule-based command-output optimization and context-compression layer that reduces unnecessary LLM context while preserving important information.

Quor

Your AI coding assistant is burning tokens on noise. Quor cuts it before it ever reaches the model. Runs entirely on your machine. No LLM, no cloud, no network call — just a deterministic rule pipeline that strips the boilerplate out of everything your assistant reads.

35.9% smaller, on average. Up to 89% smaller on the worst offenders.

Measured across a 127-case real-world benchmark suite, CI-gated on every single change — not a one-time demo number. See the numbers below, or run quor gain / quor dashboard to see your own project's real savings.

Local-only	No LLM	No cloud
No telemetry	No API keys	No file uploads
Deterministic	Fail-open	Enterprise-safe

Install — 30 seconds

pip install quor
quor init --claude
quor doctor

Requires Python 3.11+. If quor/qr isn't found on your PATH after install, run every command below as python -m quor ... instead — that's exactly what Claude Code itself already uses under the hood, so it always works.

To upgrade: pip install --upgrade quor && quor init --claude && quor doctor — hook files live outside the package, so re-running init keeps them in sync.

Why this matters

Every command your AI assistant runs — git status, a pytest run, a file read — pours straight into its context window. Most of it is boilerplate: passing tests it doesn't need to see, unchanged diff context, repeated warnings, PDF page furniture, dependency-install spam. That's tokens you're paying for, latency you're waiting on, and context-window budget your assistant isn't spending on your actual code.

The obvious fix — have the AI summarize its own output — is the wrong one: it doubles latency, doubles cost, and can silently drop the one line that mattered. Quor takes the other path: a local, rule-based pipeline that runs in milliseconds, makes the same keep/drop decision every time given the same input, and never touches the network.

command runs → Quor captures the output → rules mark each line KEEP / COMPRESS / PROTECT → noise drops → the assistant reads fewer tokens

Same command, same exit code, same side effects — only what reaches the context window changes. Failures, diffs, and tracebacks are never touched, and every compressed output links back to the full original — nothing is ever lost, just deferred until you ask for it.

The Numbers

35.9% average token reduction across Quor's own 127-case benchmark suite (CI-gated — a regression here fails the build, not just a dashboard). That average includes plenty of already-terse output with nothing left to cut; on the cases that actually have noise to remove, it goes much further:

Real command	Compression
pip install -r requirements.txt (mostly-cached dependencies)	88.8% smaller
A deeply nested Java exception stack trace	88.6% smaller
pnpm install progress noise	77.1% smaller
A large JavaScript file read through Claude Code's Read tool	75.0% smaller

By ecosystem:

Content	Compression
Java	55.6%
Config files (JSON/TOML/YAML/lockfiles)	53.9%
JavaScript	49.5%
Python packaging (pip/poetry)	47.4%
TypeScript	42.5%
Python	35.6%
CI/build logs	35.7%
Documents (PDF, DOCX, Markdown)	24.8%

Short, already-dense output compresses little — that's correct behavior, not underperformance; Quor never trims a line just to move the number (see ANTI_GOALS.md). Full breakdown in docs/BENCHMARKS.md, or run quor gain / quor dashboard for your own project's real, live numbers — always shown with an honest ±20% uncertainty band, never a bare number dressed up as exact.

Commands

quor init --claude	Install the Claude Code hook
quor doctor	Health check
quor gain	Cumulative token savings summary
quor dashboard	Live terminal view of savings for this session
quor explain <cmd>	Show what would be removed, stage by stage
quor search <query>	Semantic search across your repository
quor map / quor symbols / quor graph	Repository profile, symbol index, and dependency graph
quor validate [file]	Validate a filter config

Full reference: quor --help.

Supported

Full compression: Claude Code (Bash + Read hooks), Gemini CLI (command rewriting). Detected, integration pending upstream hook support: Codex CLI, Cursor, VS Code (GitHub Copilot agent mode), Windsurf (Cascade), Aider, Continue.dev — quor doctor tells you exactly which state you're in.

Commands: git, pytest, mypy/ruff, pip/poetry, the full Node/TypeScript toolchain (npm, pnpm, yarn, ESLint, tsc, Jest, Vitest, Prettier, Next.js, Turbo), and a generic fallback for everything else. Source code: Python built in; JavaScript/TypeScript, Go, Rust, Java, C# via pip install "quor[<language>]". Documents: Markdown, TXT, DOCX, PDF via quor[documents]. Config: JSON/TOML/.env/.ini built in, YAML via quor[yaml].

Trust

Compression tooling sits in the middle of every command you run — it has to earn the right to be there.

• Local execution only — no network calls, no cloud, no telemetry, no API keys, ever
• Rule-based, not AI — pattern match, dedup, count, budget; zero ML in the filter path, zero hallucination risk
• Fail-open — a filter bug never blocks a command or hides output, it just returns the original untouched
• Secret-aware — warns (never silently strips) if a credential pattern survives compression
• Meaning-preserving by contract — a line Quor keeps is bit-for-bit identical to the original; nothing is rephrased or summarized (see ANTI_GOALS.md)
• App-control friendly — every invocation runs through python -m quor directly, never an unsigned launcher .exe, so corporate AppLocker/Defender policies don't get in the way

Contributing

git clone https://github.com/priyanshup/Quor.git && cd Quor
pip install -e ".[dev]"
pytest tests/

CONTRIBUTING.md · SECURITY.md · CHANGELOG.md

License

Apache 2.0 — see LICENSE