What it does

BurnRate shows what you spend on AI models in one place. The web dashboard connects OpenAI and Anthropic reporting accounts and displays provider-reported usage and cost totals, with daily spend, budgets and cost trends. The site stresses that cloud reporting does not request prompts, responses or source code. For finer detail there is an optional command-line tool that runs on your machine and reads local session data from Claude Code, Cursor, OpenAI Codex, GitHub Copilot, Windsurf, Cline and Aider. It adds session, project and subagent cost breakdowns, rate limit tracking and optimization rules that suggest savings. The site notes that these figures are local estimates, not provider invoices. An optional Relay can send configured aggregate usage to a cloud workspace without storing provider keys in the cloud. The Team plan adds a shared workspace with member invitations, roles and owner-managed seat billing. BurnRate says it is not currently SOC 2 certified.

Capabilities

Combined OpenAI and Anthropic spend with budgets
Optional local CLI for session, project and subagent cost detail
46 local optimization rules with savings estimates
Rate limit tracking for Claude, Cursor and Codex
Local Relay that shares only configured aggregate usage
Free cost calculator and agent config scanner

When to use it

You want one view of OpenAI and Anthropic spend with a budget
You want to see which projects, sessions or subagents use the most tokens in your AI coding tools
You want raw code and session data to stay on your own machine
A team needs shared aggregate AI spend reporting with roles