🚀 Maximize your product's SEO. Submit to 240+ directories in 1-click with DirSubmit. Launch Now
Burnstate logo

Burnstate

Claude Code, Codex & Cursor limits, paced in your menu bar

2026-11-13

Product Introduction

  1. Definition: Burnstate is a native macOS menu bar application built in SwiftUI specifically designed for developers who code with AI agents and tools. It acts as a real-time system monitor, tracking usage limits, tokens, and cost metrics across popular AI coding environments including Claude Code, Codex, Cursor, and Vercel AI Gateway.
  2. Core Value Proposition: Burnstate bridges the gap between raw AI token usage and real-world time management. By calculating a pace forecast against fixed limit windows (session, weekly, or billing cycle), the application provides developers with predictive analytics to prevent unexpected rate-limit exhaustion, hours before it actually occurs.

Main Features

  1. AI Limit & Pace Tracking: Reads authentication tokens directly from local CLI credentials or app configurations without altering existing logins. It measures usage percentages against the clock using a visual white tick representing elapsed window time, clearly indicating whether usage is burning faster than the allowed rate.
  2. Interactive Notice Cards & CLI Control: Features a floating notification system that surfaces session updates, rate-limit warnings, and agent questions without stealing window focus. It integrates directly with Claude Code via ~/.claude/settings.json hooks, enabling developers to approve CLI bash execution prompts and answer multiple-choice questions directly from the menu bar card.
  3. Session Aggregation & Burn Reports: Consolidates all active agent sessions into a unified dashboard tracking tokens, estimated per-token costs, cache hit rates, and model distributions (such as Opus and Sonnet). It compiles up to a year of historical coding metrics into a secure, URL-encoded "burn report" that keeps all data locally on the machine.

Problems Solved

  1. Pain Point: Hitting unexpected AI rate limits mid-workflow, losing context when agents pause for tool permissions, and suffering from digital fatigue due to excessive context-switching between code editors and terminal windows.
  2. Target Audience: Software engineers, AI application developers, systems architects, and technical founders who heavily utilize autonomous terminal agents (like Claude Code and Codex) or IDE-integrated AI coding assistants (like Cursor).
  3. Use Cases: Monitoring remaining session tokens during heavy refactoring sprints; pausing work via Cooldown reminders to prevent terminal fatigue; managing remote CLI agent prompts remotely from the menu bar; and auditing monthly API spend through local log computations.

Unique Advantages

  1. Strengths & Limitations (Pros & Cons):
    • Strengths: Zero account registration required; strict local-first privacy policy with zero telemetry or tracking; native SwiftUI performance; direct loopback integration for Claude Code prompt resolution; and flexible purchase models (one-time license or subscription).
    • Limitations: Requires macOS 15 Sequoia or later; limited initially to specific AI coding tooling (Claude Code, Codex, Cursor, and Vercel AI Gateway); cost calculations rely on public API rate estimations rather than direct billing API syncs for non-API subscriptions.
  2. Key Alternatives & Differentiation: Unlike standard browser dashboard extensions or generic macOS system monitors (such as iStat Menus), Burnstate offers deep, bidirectional CLI integration (permitting/denying bash commands and answering agent prompts) combined with predictive pacing algorithms. Compared to native vendor dashboards that require constant alt-tabbing, Burnstate consolidates multi-provider telemetry into a single, unified menu bar widget with local log parsing.

Frequently Asked Questions (FAQ)

  1. Does Burnstate require an account or transmit my telemetry data? No. Burnstate operates entirely locally, reading credentials from your Mac's Keychain and computing usage, tokens, and cost statistics directly from local log files. The direct edition verifies its license key via burnstate.dev, but enforces a strict zero-analytics and zero-tracking policy.
  2. How does Burnstate handle Claude Code permission prompts? Upon initial launch, Burnstate safely inserts a hook into your ~/.claude/settings.json file. When Claude Code requests execution permission for bash commands or edits, Burnstate displays a native notification card allowing you to approve or deny the action directly from your workflow.
  3. What does the Cooldown feature do? Cooldown tracks keyboard and mouse idle times to help developers avoid burnout. It issues non-intrusive break reminders after prolonged coding sessions, locks the screen during mandatory rests (optional), allows you to queue tasks for background agents, and presents a summary of agent activity upon your return.
  4. How are cost estimates calculated? Costs are computed by applying published per-token API pricing models—accounting for cache reads, cache writes, and output tokens—directly to your local execution logs. Cursor metrics are pulled directly from Cursor application logs.

Submit to 240+ Directories with 1-Click

Maximize your product's SEO and drive massive traffic by automatically submitting it to over 240 curated startup directories using DirSubmit.

Related Products

Subscribe to Our Newsletter

Get weekly curated tool recommendations and stay updated with the latest product news