BurnRate

BurnRate is a monitoring and analytics tool focused on AI cost visibility, with the core goal of eliminating blind spots in AI model spending so users can see clearly "which model every cent was spent on."

Uses

• Aggregate costs from multiple model providers such as OpenAI and Anthropic into one dashboard to view daily spending, budget usage, and cost trends.
• Track session- and agent-level consumption of coding AI tools (such as Claude Code, Cursor, OpenAI Codex, GitHub Copilot, Windsurf, Cline, and Aider) to understand actual usage costs.
• Set budgets and alerts to be notified when budgets are at risk of overruns.

Main capabilities

• Provider report aggregation: Connect provider reporting accounts to obtain provider-reported usage and cost totals; importing aggregated usage is also supported.
• Local CLI analysis: An optional local command-line tool provides session-, project-, and agent-level analytics, including API-equivalent cost per session, cache hit rate, sub-agent cost trees, tool call tracking, operation-level token breakdowns, and more.
• Cost comparison and optimization: Provides cost comparison, efficiency scoring, an optimization recommendation engine (materials mention 46 rules, including configuration snippets for each provider), model recommendations, context efficiency analysis, and more.
• Budgets and alerts: Supports monthly and daily limits by project, burn rate monitoring and forecasting, as well as desktop notifications and Slack webhook alerts.
• Rate limit monitoring: Allows viewing 5-hour, daily, and weekly window usage for Claude, Cursor, and Codex, with color thresholds and reset countdowns.
• Team shared views: The cloud workspace supports aggregated provider reports, member invitations and roles (admin, member, read-only viewer), and seat billing managed by the owner.

How to use

1. Start using the cloud dashboard for free in a browser and choose a workspace.
2. Connect reporting sources: add provider reporting accounts, import aggregated usage, or explicitly pair and configure the optional Relay.
3. If session-, project-, and agent-level details are needed, install the optional local CLI. The materials mention that the CLI automatically parses local JSONL files with zero configuration.
4. In the Team plan, owners and admins can invite members and assign workspace roles, and manage budgets and seat billing.

Privacy by design The official statement says cloud reporting does not request prompts, responses, or source code; credentials are stored encrypted, providingThe merchant key can be kept locally (aggregated reports are sent via a local Relay, and the key is not stored in the cloud). The local CLI code and raw session data do not leave the user's machine, and core analysis has no cloud dependency. To share local data to a cloud workspace, separate pairing and Relay configuration are required; a team report stream will not be created automatically after installing the CLI.

Use cases

• Individual developers or teams want a unified view of actual spending across multiple AI model providers.
• Users of multiple AI coding tools want to compare costs, identify high-consumption sessions, and optimize usage.
• Users who need to control budgets by project, set alerts, and monitor rate limits.

Plans and limitations The information shows that the cloud dashboard is divided into three plans: Free, Pro, and Team, with an additional contact option for organizational needs. Free includes basic tracking, a cloud overview, and a provider spending ledger, with up to 3 usage sources (including 1 Relay) and 1 workspace budget. Pro includes unlimited provider and Relay sources, provider and model comparisons, explainable savings suggestions, automated Slack and email alerts, unlimited history, and more. Team includes shared team workspaces, aggregated provider reports, member invitations and roles, owner-managed seat billing, and priority support. Billing is per seat per month, and the workspace owner selects seats and reviews the order before checkout. The information emphasizes that budgets do not prevent provider spending.

Other tools The information also mentions a free-to-use AI cost calculator (to estimate monthly team spending and savings scenarios without login or data upload), an Agent configuration scanner (to check risky permissions, insecure MCP configurations, and missing approval gates, with files remaining in the browser), and resources such as the 2026 cost report.

Not yet provided: exact price figures for each plan, specific supported payment methods, complete terms of service beyond the free and paid plans, and detailed configuration steps for each provider integration.