Every tool call. Every repeated read. Every bloated response. Priced to the cent, per task, live in your terminal. Then claude-turbo blocks the waste — 60–90% gone before it ever hits your context window.
A local TUI that tails your active session and tells you, to the cent, what each task is costing you. No telemetry. No cloud. Your logs never leave your machine.
Every other Claude Code tool is a black box. claude-turbo opens the box — live, locally — and then closes the leaks.
A terminal TUI that tails your active Claude Code session. Timeline of every tool call, token breakdown per tool, waste heatmap, and cost-per-task in dollars. Export to HTML to share.
Hooks fire before the tool call runs. A blocked redundant read costs zero tokens. Other tools can only trim after the damage is done.
A local proxy strips SSE heartbeats, base64 images, and bloated tool results before they hit your context window. Self-tunes from aggregate usage data.
Most tools are static. claude-turbo has a feedback loop that tunes itself from aggregate usage data across all clients.
Deny a tool call (e.g., redundant file read)
Detect override — Claude retries same args within 60s? That's a false positive.
Record override count per heuristic in cumulative stats
Apply new model params — thresholds shift, weights change, bad heuristics get disabled
Aggregate fires + overrides across all clients
Compute FP rate per heuristic. >30% = penalize, <5% = boost.
Retune — adjust weights AND thresholds. High FP on H1? Allow 2 re-reads instead of 1.
Hook layer blocks waste upfront. Proxy layer filters the wire. Model layer tunes everything from evidence.
7 hook heuristics block waste upfront. 3 proxy filters clean the wire. All thresholds tunable from the server.
cargo test / pytest? Rewrites to scope to the active test.go build / cargo check twice? Denied. Output hasn't changed.--oneline -20. Limit tunable from server.ls -R / unbounded find? Pipes through head -50.bash_progress events dropped at the wire. Zero tokens.data:image/... payloads replaced with placeholder. Saves 10K+ tokens per image.Every heuristic was derived by analyzing actual Claude Code session logs — no guesses, no synthetic benchmarks.
Raw Claude Code session logs processed through the analyzer to find repeat patterns and payload bloat.
Saveable tokens identified across that corpus — repeated file reads, duplicate builds, unscoped tests, massive outputs.
Point claude-turbo analyze at your own ~/.claude/projects logs. See exactly where your tokens went before you install anything.
claude-turbo runs locally and intercepts tool calls on your own machine. The only thing that ever goes to our server is a handful of counters — and even that you can turn off in one line.
macos, linux, windows)Read// POST https://api.claude-turbo.xyz/v1/telemetry { "client_id": "a3f2c8d19e4b7f60c1d8e2a5f9b4c7e1", // random, local "version": "0.3.0", "platform": "macos", "days_active": 12, "total_tokens_saved": 184320, "total_interceptions": 247, "total_invocations": 1893, "heuristic_totals": { "H1_redundant_read": 94, "H4_duplicate_bash": 38, /* ... */ }, "heuristic_overrides": { "H1_redundant_read": 2, /* ... */ }, "daily": [ { "date": "2026-04-11", "tokens_saved": 12840, /* ... */ } ], "proxy_running": true }
~/.claude-turbo/config.json:
"telemetry_enabled": false
"telemetry_url": "https://your.host"
github.com/blasrodri/claude-turbo
One binary. Dashboard + proxy + heuristics. No config. Gets smarter every session.
curl -fsSL https://claude-turbo.xyz/install.sh | sh