Observed arrival · 2026-10-26
Tokenmetry: AI coding spend, gathered locally
A local-first desktop dashboard that reads usage logs from 19 AI coding agents and organizes token use, costs, sessions, projects, and activity into fixed views.
- For
- People tracking usage across multiple AI coding agents
- Worth noticing
- Unknown-priced models remain marked unknown rather than being counted as zero; the page lists only price refreshes and update checks as network activity.
Field notes
Its readers access JSONL, JSON, compressed session logs, or SQLite databases in read-only mode; changed files are incrementally rescanned by modification time and size, then normalized into a local SQLite ledger. Nine fixed views cover sessions through pricing, while the tray can show today's tokens, estimated cost, cache hits, and rate. The page says prompts, code, and session contents are not stored; its listed network touches are background price refreshes and update checks. Public installers are pending, and source use currently requires the .NET 10 SDK.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue