Summer tracks AI coding usage, costs

Tracking developer AI spend is becoming a necessity as teams adopt high-volume agents like Claude Code that can quickly rack up thousands in API costs.

• –Provides visibility into which engineers, models, and tools are driving the most costs, helping organizations optimize their AI tool investments.
• –Offers a zero-hosting architecture by utilizing a local daemon to collect events and using Autumn's API as the backend for token pricing (via Models.dev) and aggregation.
• –Includes a local browser dashboard that allows teams to group, filter, and search usage data by harness, model, user, or billing mode.
• –Requires a restart of active Claude Code sessions after setup to begin live tracking, which may add slight friction to initial onboarding.