hatch3r-cost-tracking

To estimate token usage and cost when metrics are unavailable, calculate estimated tokens per content type and apply model-tier pricing. This generates a structured cost report and enforces session, issue, and epic-level budgets.

Budget guardrails prevent surprise AI spend by comparing estimated token spend against configured thresholds. When limits are crossed, the system triggers warnings, hard stops, and recommends concrete optimizations to reduce context bloat.

Track AI spend across session, issue, and epic levels by estimating input and output tokens, converting them to dollar costs using model-tier rates, and comparing the results against your configured budgets to produce a structured cost report.

Optimize token usage by detecting common waste patterns like large reads, context bloat, batching mistakes, and scope creep. The system recommends concrete reductions and suggests how to defer or re-scope work to stay within budget.

Yes, cost tracking works with hatch.json budgets by comparing estimated spend against those configured thresholds. It triggers warnings or hard stops based on the limits defined in your configuration file to enforce budget-aware guardrails.