High-performance CLI proxy that cuts up to 90% of the bash output your agent reads
Website • Install • Troubleshooting • Architecture • Discord
English • Francais • 中文 • 日本語 • 한국어 • Espanol • Português
rtk filters and compresses command outputs before they reach your LLM context. Single Rust binary, 100+ supported commands, <10ms overhead.
RTK intercepts shell commands and compresses their output before your agent reads it.
| Operation | What RTK does to the output |
|---|---|
ls / tree |
Tree format with file counts instead of one line per entry |
cat / read |
Smart file reading: signatures and structure over full bodies |
grep / rg |
Truncates long lines, groups matches by file |
git status |
Compact stat format, grouped by state |
git diff |
Reduced context, headers stripped |
git log |
Hash, author and subject only |
git add/commit/push |
Confirmation line instead of full progress output |
cargo test / npm test |
Failures only, passing tests collapsed to a count |
ruff check |
Grouped by rule and file |
pytest |
Failures only, traceback trimmed |
go test |
NDJSON parsed, failures only |
docker ps |
Essential fields only |
RTK cuts up to 90% of the bash output your agent reads. That is what RTK measures, and it is not the same as cutting your bill by 90%.
Bash output is one contributor to input tokens, alongside your prompt, the system prompt and conversation history. Input tokens are in turn only part of the bill, which also counts output tokens. The reduction dilutes at every step.
The token counts RTK reports are estimated as bytes / 4 — RTK ships no tokenizer, so the percentages are reliable but the absolute token numbers are approximate.
Full explanation: How RTK Savings Work
brew install rtkcurl -fsSL https://raw.githubusercontent.com/rtk-ai/rtk/refs/heads/master/install.sh | shInstalls to
~/.local/bin. Add to PATH if needed:echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc # or ~/.zshrc
cargo install --git https://github.com/rtk-ai/rtkDownload from releases:
- macOS:
rtk-x86_64-apple-darwin.tar.gz/rtk-aarch64-apple-darwin.tar.gz - Linux:
rtk-x86_64-unknown-linux-musl.tar.gz/rtk-aarch64-unknown-linux-gnu.tar.gz - Windows:
rtk-x86_64-pc-windows-msvc.zip
Windows users: Extract the zip and place
rtk.exesomewhere in your PATH (e.g.C:\Users\<you>\.local\bin). Run RTK from Command Prompt, PowerShell, or Windows Terminal — do not double-click the.exe(it will flash and close). The full hook system works natively on Windows (and in WSL). See Windows setup below for details.
rtk --version # Should show "rtk 0.28.2"
rtk gain # Should show the savings dashboardName collision warning: Another project named "rtk" (Rust Type Kit) exists on crates.io. If
rtk gainfails, you have the wrong package. Usecargo install --gitabove instead.
# 1. Install for your AI tool
rtk init -g # Claude Code / Copilot (default)
rtk init -g --gemini # Gemini CLI
rtk init -g --codex # Codex (OpenAI)
rtk init -g --agent cursor # Cursor
rtk init -g --agent windsurf # Windsurf
rtk init --agent cline # Cline / Roo Code
rtk init --agent kilocode # Kilo Code
rtk init --agent antigravity # Google Antigravity
rtk init --agent kimi # Kimi AI
rtk init -g --agent pi # Pi
rtk init --agent hermes # Hermes
rtk init -g --agent droid # Factory Droid
# 2. Restart your AI tool, then test
git status # Automatically rewritten to rtk git statusHook-based agents rewrite Bash commands (e.g., git status -> rtk git status) before execution. Plugin-based agents, including Hermes, use their plugin API to rewrite commands before execution. The agent receives compact output without needing to call rtk explicitly.
Important: the hook only runs on Bash tool calls. Claude Code built-in tools like Read, Grep, and Glob do not pass through the Bash hook, so they are not auto-rewritten. To get RTK's compact output for those workflows, use shell commands (cat/head/tail, rg/grep, find) or call rtk read, rtk grep, or rtk find directly.
Without rtk: With rtk:
Claude --git status--> shell --> git Claude --git status--> RTK --> git
^ | ^ | |
| full raw output | | compact output | filter |
+-----------------------------------+ +------- (filtered) ---+----------+
Four strategies applied per command type:
- Smart Filtering - Removes noise (comments, whitespace, boilerplate)
- Grouping - Aggregates similar items (files by directory, errors by type)
- Truncation - Keeps relevant context, cuts redundancy
- Deduplication - Collapses repeated log lines with counts
Percentages below are reductions in bash output, not reductions in your bill. See How Savings Work.
rtk ls . # Compact directory tree
rtk read file.rs # Smart file reading
rtk read file.rs -l aggressive # Signatures only (strips bodies)
rtk smart file.rs # 2-line heuristic code summary
rtk find "*.rs" . # Compact find results
rtk grep "pattern" . # Grouped search results
rtk diff file1 file2 # Condensed diff (exit 1 if files differ)rtk git status # Compact status
rtk git log -n 10 # One-line commits
rtk git diff # Condensed diff
rtk git add # -> "ok"
rtk git commit -m "msg" # -> "ok abc1234"
rtk git push # -> "ok main"
rtk git pull # -> "ok 3 files +10 -2"rtk gh pr list # Compact PR listing
rtk gh pr view 42 # PR details + checks
rtk gh issue list # Compact issue listing
rtk gh run list # Workflow run statusrtk jest # Jest compact (failures only)
rtk vitest # Vitest compact (failures only)
rtk playwright test # E2E results (failures only)
rtk pytest # Python tests (-90%)
rtk go test # Go tests (NDJSON, -90%)
rtk cargo test # Cargo tests (-90%)
rtk rake test # Ruby minitest (-90%)
rtk rspec # RSpec tests (JSON, -60%+)
rtk err <cmd> # Filter errors only from any command
rtk test <cmd> # Generic test wrapper - failures only (-90%)rtk lint # ESLint grouped by rule/file
rtk lint biome # Supports other linters
rtk tsc # TypeScript errors grouped by file
rtk next build # Next.js build compact
rtk prettier --check . # Files needing formatting
rtk cargo build # Cargo build (-80%)
rtk cargo clippy # Cargo clippy (-80%)
rtk ruff check # Python linting (JSON, -80%)
rtk golangci-lint run # Go linting (JSON, -85%)
rtk rubocop # Ruby linting (JSON, -60%+)
rtk sbt test # ScalaTest output (-90%)
rtk sbt compile # Compilation errors only (-75%)
rtk sbt run # Strip SBT preamble noisertk pnpm list # Compact dependency tree
rtk uv run pytest # Preserve uv env, keep program output
rtk pip list # Python packages (auto-detect uv)
rtk pip outdated # Outdated packages
rtk bundle install # Ruby gems (strip Using lines)
rtk prisma generate # Schema generation (no ASCII art)rtk aws sts get-caller-identity # One-line identity
rtk aws ec2 describe-instances # Compact instance list
rtk aws lambda list-functions # Name/runtime/memory (strips secrets)
rtk aws logs get-log-events # Timestamped messages only
rtk aws cloudformation describe-stack-events # Failures first
rtk aws dynamodb scan # Unwraps type annotations
rtk aws iam list-roles # Strips policy documents
rtk aws s3 ls # Truncated with tee recoveryrtk docker ps # Compact container list
rtk docker images # Compact image list
rtk docker logs <container> # Deduplicated logs
rtk docker compose ps # Compose services
rtk kubectl pods # Compact pod list
rtk kubectl logs <pod> # Deduplicated logs
rtk kubectl services # Compact service list
rtk oc get pods # OpenShift pod summary
rtk oc get services # OpenShift service list
rtk oc logs <pod> # Deduplicated logsrtk pulumi preview # Strip header/URL/duration noise
rtk pulumi up # Compact apply output
rtk pulumi destroy # Compact destroy output
rtk pulumi refresh # Drift summary
rtk pulumi stack # Stack metadata (strips owner/timestamps)rtk json config.json # Structure without values
rtk deps # Dependencies summary
rtk env -f AWS # Filtered env vars
rtk log app.log # Deduplicated logs
rtk curl <url> # Truncate + save full output
rtk wget <url> # Download, strip progress bars
rtk summary <long command> # Heuristic summary
rtk proxy <command> # Raw passthrough + trackingrtk gain # Summary stats
rtk gain --graph # ASCII graph (last 30 days)
rtk gain --history # Recent command history
rtk gain --daily # Day-by-day breakdown
rtk gain --all --format json # JSON export for dashboards
rtk discover # Find missed savings opportunities
rtk discover --all --since 7 # All projects, last 7 days
rtk session # Show RTK adoption across recent sessions-u, --ultra-compact # ASCII icons, inline format (further output reduction)
-v, --verbose # Increase verbosity (-v, -vv, -vvv)Directory listing:
# ls -la (45 lines) # rtk ls (12 lines)
drwxr-xr-x 15 user staff 480 ... my-project/
-rw-r--r-- 1 user staff 1234 ... +-- src/ (8 files)
... | +-- main.rs
+-- Cargo.toml
Git operations:
# git push (15 lines) # rtk git push (1 line)
Enumerating objects: 5, done. ok main
Counting objects: 100% (5/5), done.
Delta compression using up to 8 threads
...
Test output:
# cargo test (200+ lines on failure) # rtk test cargo test (~20 lines)
running 15 tests FAILED: 2/15 tests
test utils::test_parse ... ok test_edge_case: assertion failed
test utils::test_format ... ok test_overflow: panic at utils.rs:18
...
The most effective way to use rtk. The hook transparently intercepts Bash commands and rewrites them to rtk equivalents before execution.
Result: 100% rtk adoption across all conversations and subagents, with no per-command context overhead.
Scope note: this only applies to Bash tool calls. Claude Code built-in tools such as Read, Grep, and Glob bypass the hook, so use shell commands or explicit rtk commands when you want RTK filtering there.
rtk init -g # Install hook + RTK.md (recommended)
rtk init -g --opencode # OpenCode plugin (instead of Claude Code)
rtk init -g --auto-patch # Non-interactive (CI/CD)
rtk init -g --hook-only # Hook only, no RTK.md
rtk init --show # Verify installationAfter install, restart Claude Code.
RTK works fully on native Windows. Since v0.37.2 the auto-rewrite hook runs as a native binary command (rtk hook claude) — no Unix shell, bash, or jq required — so commands are rewritten transparently on Command Prompt, PowerShell, and Windows Terminal, just like on Linux and macOS.
# 1. Download and extract rtk-x86_64-pc-windows-msvc.zip from releases
# 2. Add rtk.exe to your PATH (e.g. C:\Users\<you>\.local\bin)
# 3. Initialize — installs the native binary hook
rtk init -gUpgrading from an older install? If you set RTK up before v0.37.2 you may still have the legacy rtk-rewrite.sh shell hook (which does need a Unix shell). Re-run rtk init -g to migrate to the native binary hook.
Prerequisites: some filters shell out to ripgrep (rg). Install it and keep it on your PATH (e.g. winget install BurntSushi.ripgrep.MSVC) to avoid Binary 'rg' not found on PATH warnings.
Important: Do not double-click rtk.exe — it is a CLI tool that prints usage and exits immediately. Always run it from a terminal (Command Prompt, PowerShell, or Windows Terminal).
WSL also works and behaves exactly like Linux:
# Inside WSL
curl -fsSL https://raw.githubusercontent.com/rtk-ai/rtk/refs/heads/master/install.sh | sh
rtk init -g| Feature | Native Windows | WSL |
|---|---|---|
| Filters (cargo, git, etc.) | Full | Full |
| Auto-rewrite hook | Yes (native binary) | Yes |
rtk init -g |
Hook mode | Hook mode |
rtk gain / analytics |
Full | Full |
RTK supports 16 AI coding tools. Each integration rewrites shell commands to rtk equivalents, reducing the bash output the agent reads where the agent supports command interception.
| Tool | Install | Method |
|---|---|---|
| Claude Code | rtk init -g |
PreToolUse hook (native binary) |
| GitHub Copilot (VS Code) | rtk init -g --copilot |
PreToolUse hook — transparent rewrite |
| GitHub Copilot CLI | rtk init -g --copilot |
PreToolUse deny-with-suggestion (CLI limitation) |
| Cursor | rtk init -g --agent cursor |
preToolUse hook (hooks.json) |
| Gemini CLI | rtk init -g --gemini |
BeforeTool hook |
| Codex | rtk init -g --codex |
AGENTS.md + RTK.md instructions |
| Windsurf | rtk init -g --agent windsurf |
.windsurfrules (project-scoped) |
| Cline / Roo Code | rtk init --agent cline |
.clinerules (project-scoped) |
| OpenCode | rtk init -g --opencode |
Plugin TS (tool.execute.before) |
| OpenClaw | openclaw plugins install ./openclaw |
Plugin TS (before_tool_call) |
| Pi | rtk init -g --agent pi (global) |
TypeScript extension (tool_call) |
| Hermes | rtk init --agent hermes |
Python plugin adapter (terminal command mutation via rtk rewrite) |
| Mistral Vibe | rtk init -g --agent vibe |
pre_tool hook (hooks.toml) |
| Kilo Code | rtk init --agent kilocode |
.kilocode/rules/rtk-rules.md (project-scoped) |
| Google Antigravity | rtk init --agent antigravity |
.agents/rules/antigravity-rtk-rules.md (project-scoped) |
| Kimi AI | rtk init --agent kimi |
AGENTS.md (project-scoped) |
| Factory Droid | rtk init -g --agent droid (or per-project) |
PreToolUse hook in ~/.factory/hooks.json (matcher Execute) |
For per-agent setup details, override controls, and graceful degradation, see the Supported Agents guide. The Hermes plugin source and tests live in hooks/hermes/; installed Hermes runtime files still live under ~/.hermes/plugins/rtk-rewrite/.
~/.config/rtk/config.toml (macOS: ~/Library/Application Support/rtk/config.toml):
[hooks]
exclude_commands = ["curl", "playwright"] # skip rewrite for these
[tee]
enabled = true # save raw output on failure (default: true)
mode = "failures" # "failures", "always", or "never"When a command fails, RTK saves the full unfiltered output so the LLM can read it without re-executing:
FAILED: 2/15 tests
[full output: ~/.local/share/rtk/tee/1707753600_cargo_test.log]
For the full config reference (all sections, env vars, per-project filters), see the Configuration guide.
rtk init -g --uninstall # Remove hook, RTK.md, settings.json entry
cargo uninstall rtk # Remove binary
brew uninstall rtk # If installed via Homebrew- rtk-ai.app/guide — full user guide (installation, supported agents, what gets optimized, analytics, configuration, troubleshooting)
- INSTALL.md — detailed installation reference
- ARCHITECTURE.md — system design and technical decisions
- CONTRIBUTING.md — contribution guide
- SECURITY.md — security policy
RTK sends no telemetry and never opens network connections on its own. Commands requested by you, such as rtk curl, rtk wget, git push, or docker, may use the network; set RTK_OFFLINE=1 to reject RTK's curl and wget proxies, and use an OS sandbox when child-process network isolation is required.
Local tracking is disabled by default. When enabled, it stores only daily aggregate counts by broad category plus token and duration totals—never command lines, arguments, paths, output, errors, device identifiers, or user identifiers. Retention defaults to 30 days and is controlled by tracking.history_days.
[tracking]
enabled = false
history_days = 30To permanently remove a legacy detailed history database (including SQLite WAL/SHM sidecars), run rtk privacy purge-history --yes. Hook changes support --dry-run and list the files they would modify.
- Patrick Szymkowiak — Founder GitHub · LinkedIn
- Florian Bruniaux — Core contributor GitHub · LinkedIn
- Adrien Eppling — Core contributor GitHub · LinkedIn
- Nicolas Le Cam — Core contributor Github · LinkedIn
- Takayuki Maeda — Core contributor GitHub · LinkedIn
Contributions welcome! Please open an issue or PR on GitHub.
Join the community on Discord.
Apache License 2.0 - see LICENSE for details.
See DISCLAIMER.md.