BurnRate

BurnRate is a local-first observability and cost analytics tool built for developers to track and optimize spending on AI coding assistants. It runs directly from the terminal with zero cloud dependencies, parsing local JSONL automatically to keep user data entirely on the machine. The software provides nine analytics views, subagent tracking, and multi-provider rate limit monitoring to help developers understand where tokens go and control their expenses.

• Subagent observability to track every subagent Claude Code spawns, including its cost, duration, depth, and tool usage.
• Multi-provider analytics supporting Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Cline, and Aider.
• An optimization engine featuring over 40 rules that analyze usage patterns and offer one-click configuration snippets.
• Project analytics that track costs per codebase, monitor budget caps, and rank projects by expense.
• Operation-level breakdown measuring token consumption across reading, writing, searching, and thinking operations.
• Rate limit monitoring with live utilization bars, color-coded thresholds, and reset countdowns across providers.
• Budget alerts that trigger desktop notifications or Slack webhooks when monthly and daily caps are at risk.

Pros

• Local-first architecture ensures that user code and session data never leave the machine, avoiding cloud data dependencies.
• Deep subagent visibility in Claude Code reveals token distribution across complex agentic workflows.
• Actionable optimization rules provide concrete savings estimates and provider-specific configuration snippets.
• Multi-provider comparison aggregates usage across seven different coding tools into a single dashboard.

Cons

• The free tier restricts history retention to 30 days and limits support to three providers.
• Advanced optimization engine features, model recommendations, and JSON export require upgrading to a paid plan.
• Team dashboards, aggregate analytics, and admin consoles are locked behind the Team tier pricing.
• The page does not detail enterprise-specific feature limits or custom deployment requirements for the Enterprise tier.

BurnRate is a strong fit for individual developers and engineering teams heavily utilizing AI coding assistants who need to audit token usage, prevent unexpected rate limits, and manage spending. It is well-suited for developers who prioritize data privacy and require local execution without cloud telemetry. It is a poor fit for organizations that do not use supported AI coding tools or those looking for an entirely cloud-hosted solution without local command-line setup.

BurnRate uses a freemium pricing model with four tiers. The Free tier costs $0 forever and includes basic analytics for 3 provider integrations with 30-day history retention. The Pro tier costs $12 per month, offering unlimited history, all nine analytics tabs, optimization engine access, model recommendations, and JSON exports. The Team tier costs $29 per seat per month (with a baseline of $145 per month for 5 seats) and adds team dashboards, aggregate analytics, an admin console, usage benchmarking, and priority support. An Enterprise tier is also listed, though specific pricing for it is not published on the page.

Unlike generic cloud-based cost management tools, BurnRate operates entirely locally on the user's machine with zero cloud dependencies, parsing local JSONL files directly. It uniquely surfaces subagent behavior and cost trees for Claude Code workflows, addressing closed feature requests where other tools cannot see inside spawned subagents. Furthermore, it combines multi-provider rate limit tracking with an action-oriented optimization engine that automatically generates configuration snippets to cut costs.

Claude Code, Cursor, OpenAI Codex, GitHub Copilot, Windsurf, Cline, Aider, and Slack for alerts.