🛠 Wattage: Controlling Token Costs for AI Agents

Wattage has been released — a tool for profiling token expenses and controlling cost regressions in AI agents. It analyzes OpenTelemetry (OTLP) traces, identifies inefficient patterns, and calculates their cost in dollars. It allows for the implementation of a "cost gateway" in CI/CD.

🌍 It shifts AI agent cost management from intuitive to engineering-based using metrics.

👤 It helps find budget "leaks" when using LLMs and protects against regressions when updating prompts.

Source 1: https://github.com/faizannraza/wattage