Cut your AI coding bill by 30–60%.Without changing how your team works.

Same dev tools.
Keep using Claude Code, Codex, etc.
No new harnesses.
Our sidecar optimizes spend fully in the background.
One-time, 5-minute setup.
Homebrew install and 5-step CLI init.

Start your free trial See the live demo

21 days, no credit card 5-minute setup Track is free forever

Optimize Savingsacme.dev · this month

Sample data

Est. savings (API-rate value)$3.2K–$4.1Kvs your team's usage before Code Lexica

Tokens / message−31%1,250 with · 1,800 without

Code Lexica adoption9/12devs using Code Lexica this window

Monthly spend, same team

Model mix −$1.1KInput tokens / msg −$1.6KOutput tokens / msg −$0.9KCache −$0.6K

Weekly AI Usage BriefingJul 6 – Jul 12

Emailed

AI usage rose 13% this week, led by the checkout-service payments push. The team pulled $1,015 of API-rate value from its plans, on pace for 2.4× what the seats cost this month. Worth watching: Claude Code is now 62% of True Cost, up from 48%.

checkout-service True Cost +34% — Opus 4.8 on a large refactor

alex@acme.dev drove 21% of org tokens (mostly Opus)

Cursor usage flat; Claude Code up 18% week-over-week

LLM Usage Dashboardacme.dev · last 30 days

All tools

Tokens (30d)418M

API-rate value$4,350

You paid$1,800

Extracted value+$2,550(2.4×)

Claude CodeCodexCursor

Team Leaderboardacme.dev · last 30 days

Shared org-wide

Together: 418M tokens · $4,350 of API-rate value

You paid $1,800 · extracted +$2,550 (2.4×) beyond it

Priya Sharma

75.2Mtok

9-day streak

+$583(3.9×)

Alex Rivera

87.8Mtok

12-day streak

+$713(4.6×)

Sam Chen

58.5Mtok

5-day streak

+$409(3.0×)

Optimize savings

Works with the tools your team already uses

Claude Code

Codex

Cursor

GitHub CopilotSoon

Gemini CLISoon

The utilization problem

Every AI seat is a bet. Most orgs can't see the payoff.

You rolled out Max plans, Cursor seats, Copilot licenses. Then the economics shifted under you: Anthropic moved programmatic usage to full API rates, and Copilot metered agent mode. Meanwhile nobody can answer the basic questions. Who's actually using their seat? Which ones sit idle? Are your heaviest users throttled by limits mid-week?

The stakes compound because agent usage is enormous, and most of those tokens go to waste. The teams that thrive next won't have the biggest budgets. They'll know exactly what they're spending, and they'll cut it without slowing anyone down.

Start by seeing where it goes

~70%

of agent tokens are waste: re-grep'ing, re-reading, exploring dead ends. (Stanford agent-trace research)

$10–$15

API tokens burned per complex SWE-Bench Pro task on Opus 4.8.

25×

effective cost increase for teams running claude -p on subsidized Max plans after the June 2026 shift.

OptimizeAvailable now

Same plans. Smaller bill.

Optimize runs as a lightweight sidecar next to each developer's coding agent, steering model, effort, and context decisions in the background. Same engineers, same tools, fewer tokens. See how Optimize works.

Optimize (Standard)

up to 30%

nothing indexed, no code leaves your machines

Optimize (Premium)

30–60%

your codebase indexed on Code Lexica

Try Optimize free for 21 days

No credit card. Reverts to free Track automatically. And if we don't save you money, you don't pay.

Start your free trial See pricing

Proof

Two ways to prove the savings

One-off · Controlled

Paired A/B test

A controlled experiment on your own repository, finished in an afternoon.

cdlx ab-test run --task "…"

Same task, same base commit, two isolated worktrees: one arm with Code Lexica, one without. The scorecard reports tokens, cost, duration, and verify-command pass rate per arm, so the savings never hide a quality drop.

We run the experiment on your repository, not on a benchmark.

Ongoing · Real-world

Ongoing Savings Tracker

Sample Data

Live savings on your team's real work, measured against your own pre-Code-Lexica baseline.

Est. savings (API-rate value)

$3.2K–$4.1K

vs your team's usage before Code Lexica

Tokens / message

−31%

1,250 with · 1,800 without

Code Lexica adoption

9/12

devs using Code Lexica this window

More on how we measure — including what the sample above looks like on your own data.

How it works

Tracking in 5 minutes. Proof in an afternoon.

A few commands in a terminal. You won't paste an API key or route through a proxy, and nothing changes in anyone's workflow.

Install

Install the Code Lexica CLI tool via Homebrew.

brew install codelexica/tap/cdlx

Track

Connect to your Code Lexica account and select repos to track.

cdlx init

Optimize

Enable token optimization to run in the background.

cdlx sidecar enable

Measure

Measure real savings live on actual coding tasks in your codebase.

cdlx ab-test run

Start your free trial

TrackFree forever

Not ready to optimize? See the bill first.

Track is the free half of Code Lexica: real usage and True Cost for your whole team, with nothing to pay and nothing indexed. It's also the baseline Optimize measures savings against, so starting free is starting the clock.

Usage and True Cost across Claude Code, Codex, and Cursor, in one team-level view
Weekly briefings that explain what changed and why
A leaderboard you can share org-wide, no accounts needed
Unlimited team members, free forever

Explore Track See the demo leaderboard

Team Leaderboardacme.dev · last 30 days

Shared org-wide

Together: 418M tokens · $4,350 of API-rate value

You paid $1,800 · extracted +$2,550 (2.4×) beyond it

Priya Sharma

75.2Mtok

9-day streak

+$583(3.9×)

Alex Rivera

87.8Mtok

12-day streak

+$713(4.6×)

Sam Chen

58.5Mtok

5-day streak

+$409(3.0×)

4JLJordan Lee50.2M+$322(2.6×)

5?Anonymous46.0M+$278(2.4×)

Teams on Track

Why teams keep it running

“Most tools will tell you how many tokens you're consuming. What's been more valuable for us is understanding why usage changed from one week to the next. The insights make the product really useful.”

Tom DeMichele

CEO, DogMedia Solutions

“As model pricing and usage continue to change, we wanted a better understanding of our exposure. Track gave us a clearer picture of how AI was being used across teams and where costs were trending.”

Matt Blizel

Founder, Atlevo

See the bill. Then cut it.

Track is free forever, and live in 5 minutes. Optimize is free for 21 days, and if we don't save you money, you don't pay.

Start your free trial Book a demo