Early launch · Claude Code · Codex · opencode · API-backed agents

Pace your AI coding agents before you hit the wall.

Token Pacer sits in your Mac menu bar and helps you pace Claude Code, Codex, opencode, and API-backed agents across limits, usage, and spend. Try Pro for 3 days — no account, no credit card, no telemetry.

Download for Mac See pricing →

3-day Pro trial·No account·No credit card·macOS 13+

Local-first · Your logs stay on your Mac

Best Next Action

Claude 7dresets in 2d 9h

Codex secondaryresets in 3d 17h

Claude Design86% · elevated

Weekly runwayon pace

The problem

AI coding agents make it easy to sprint without noticing.

One more prompt. One more refactor. One more model switch. Then suddenly you hit a limit, burn API credits, or realize a single session consumed more than expected.

Subscription pressure

Limitscreeping up

A long coding session pushes Claude or Codex closer to a rolling limit window.

API spend rising

Spendquietly stacking

API-backed agents add cost in the background while you stay in flow.

Long session spike

Spikeone-off session

One refactor or autonomous run consumes more tokens than the rest of the day.

Reset timing

Waitor switch agents?

Know whether to wait, switch agents, or continue lightly.

The promise

Keep your AI workflow on pace.

Three signals, one decision. Token Pacer turns scattered logs and provider windows into a calm, single-screen answer to the only question that matters.

Today's activity

See what is happening now.

Limit pressure

Know when quota windows or API budgets are getting tight.

Best next action

Decide whether to keep going, switch agents, slow down, or wait.

Feature 01 — Best Next Action

Know what to do next.

The main tab turns scattered usage signals into a clear recommendation: keep going, switch agents, use API fallback lightly, wait for reset, or avoid long runs.

Keep using Claude Use Codex Use opencode lightly Wait for reset Avoid long runs

Feature 02 — Limit Windows

See rolling limits before they interrupt you.

Track Claude, Codex, and provider-native windows with reset timing and pressure clearly labeled. Use provider-native signals where available — local estimates only when needed, and always labeled as such.

Provider-reported windows where available, including Claude and Codex rolling limits.
Reset pressure at a glance — red for blocked, orange for elevated, green for healthy.
Source-of-truth labels. Provider-native, local estimate, or pricing estimate — never blurred together.

Feature 03 — Weekly Runway

Know if today's pace can last the week.

Token Pacer projects whether your current usage is sustainable across API spend and subscription limit windows — so Friday doesn't surprise you on Tuesday afternoon.

7-day pacing · Tue 4:26pm on pace

MonTueWedThuFriSatSun

Smart pacing learns your typical working hours from local history — not a generic budget.
Calm projections across subscription windows and API caps, with explicit "covered by plan" status.

Feature 04 — Recent Sessions

Find what caused the spike.

See recent agent sessions, models, projects, turns, and usage — so a long-running session doesn't stay invisible until the bill arrives or the limit lands.

Per-session breakdown by agent, model, and project — sortable by tokens or spend.
Model split — see whether kimi-k2.6, glm-5.1, or another model is doing the heavy lifting.
Active session badge highlights what's running right now, with last-update time.

Local-first

Your agent logs stay on your Mac.

No accounts. No cloud sync. No telemetry. Your agent logs stay on your Mac.

Reads local logs
Stores usage locally
No telemetry
No cloud sync
No prompt/code upload
No account required
Clear source-of-truth labels
Open Settings to inspect every source

Token Pacer watches supported local agent logs, normalizes them into usage events, and builds pacing views entirely on your Mac. It only contacts Polar when you buy, activate, refresh, or deactivate a license. Your local agent logs, project names, usage data, and pricing data are not sent to Polar.

Claude Code logs local

Codex logs local

opencode logs local

↓

Token Pacer · processes locally · stays on this Mac

logs stay local · license checks only

Supported agents

Built for multi-agent workflows.

Pace across the agents you actually use today, with a clear roadmap for adding more as reliable local signals become available.

Claude Code

Subscription windows, usage pressure, reset timing, and local session activity. Provider-native limit windows where available.

5h · 7d windowssupported

Codex

Usage windows, session activity, plan pressure, and primary/secondary window tracking where the provider exposes it.

primary · secondarysupported

opencode & API

API spend, token usage, model costs, and budget pacing across API-backed agents and inference providers.

budget · editable pricingsupported

Supports Claude Code, Codex, and opencode first. More agents will be added as reliable local signals become available.

How it works

Local signals. Clear pacing. No cloud required.

Three steps from disk to decision — same architecture, whether you use one agent or four.

Watch supported local logs

Token Pacer monitors agent logs already on your Mac — Claude Code, Codex, opencode — without ever sending them off-device.

Normalize into usage events

Each event is the source of truth. Sessions, projects, model splits, and rollups are derived from local usage data.

Show today, week, and next action

Today's activity, limit windows, weekly runway, and the best next action — labeled by source and ready to act on.

Source of truth: Usage events are the lowest-level record. Sessions, projects, and rollups are all derived from local usage data — so when something looks off, you can trace it back to a single event in a single log file.

Made for flow

Built to stay out of your way.

Token Pacer lives in your menu bar, but it does not ask you to babysit another dashboard.

Smart notifications

A quiet heads-up when limits, runway, or API spend actually need attention — not for every blip.

Keyboard shortcut

Open your current pace instantly — check, decide, get back to coding without breaking flow.

⌥⌘P

A small glyph and a single number. Click for the full picture; otherwise it stays out of the way.

Pricing

Try Pro first. Buy once if it fits.

Download Token Pacer and start a 3-day Pro trial locally. No account or credit card required.

Early launch

Token Pacer Pro

$29 one-time license 3-day Pro trial included Regular price $49 after launch

Try Pro free for 3 days. No account or credit card required.

Free modeAfter trial

Keeps basic tracking available after the Pro trial ends.

Detect supported agents
Verify local sources
Track today’s usage
Basic breakdowns

Pro license — $29Recommended

Unlimited local history
Budget guardrails
Smart pacing recommendations
CSV export
Custom model pricing
Advanced breakdowns
Smart notifications
Keyboard shortcut
3 Macs per license
All 1.x updates included

Download and start trial Buy Pro license

After the trial, Token Pacer Free still tracks today’s usage. Upgrade to Pro to keep unlimited history, budgets, exports, custom pricing, and smarter pacing.

How the trial works

Local trial. Simple license.

Download Token Pacer

Start a 3-day Pro trial automatically. No account or credit card required.

Use it with your real local logs

Token Pacer reads supported agent logs on your Mac and shows usage, pacing, and spend locally.

Keep Pro with a license

After the trial, buy a one-time license through Polar and paste the license key into the app.

Free mode stays available

If you do not upgrade, Token Pacer still tracks today’s usage with basic breakdowns.

FAQ

The honest answers.

Does Token Pacer upload my logs?

No. Agent logs, project names, usage data, and pricing data stay on your Mac. Token Pacer only contacts Polar for license purchase, activation, refresh, or deactivation.

Do I need an account to try Token Pacer?

No. The 3-day Pro trial starts locally when you open the app.

Do I need a credit card for the trial?

No. Download the app and try Pro for 3 days without checkout.

What happens after the trial?

Token Pacer switches to Free mode. You can keep basic tracking or buy a one-time Pro license.

Which agents are supported?

Claude Code, Codex, and opencode first. More sources will be added when reliable local signals are available — we'd rather support fewer agents well than many poorly.

Is API spend exact?

Sometimes. If cost is recorded in the source event, Token Pacer uses that directly. Otherwise it estimates from a pricing table you can edit, and labels the result as an estimate.

Does it track Claude/Codex subscription limits exactly?

Where provider-native signals are available, Token Pacer uses them. Otherwise it shows local estimates and labels them clearly so you always know what's measured and what's inferred.

What macOS version is required?

Token Pacer requires macOS 13 Ventura or later. It is optimized for macOS 14+ on Apple silicon.

Is this for teams?

Not yet. Token Pacer is currently designed as a single-player developer utility.

How does the license work?

After purchase, Polar sends you a license key. Paste it into Token Pacer to unlock Pro.

How many Macs can I use?

One license supports up to 3 Macs.