Read OpenCode activity from this machine
OpenCode stores tokens, requests, models, and sessions locally. An OpenCode usage tracker is useful when you want that activity in a dashboard instead of reconstructing it from session files or a CLI report.
UsageAtlas reads the local OpenCode database for per-step activity and turns it into the same tokens, requests, models, and session picture it uses for other sources. You can inspect a day, a range, or the full collected history.
Estimated cost when OpenCode does not bill you a line item
OpenCode often sits in front of several model providers, so a single invoice does not always describe the work. UsageAtlas estimates cost from token counts and bundled pricing when the source does not report a cost of its own, and labels those figures as estimates.
Use the relevant provider bill when a purchasing decision has to be exact. The dashboard is for planning: which models you actually called, and how that load compares with yesterday.
OpenCode in the same total as Codex, Claude, and Cursor
If OpenCode is one tool among several, its local logs will not tell you what Cursor or Claude spent the same afternoon. UsageAtlas adds supported OpenCode activity to one running total, with a filter when you need OpenCode alone.
Nothing extra to sign into. Prompts and responses are not read. You can switch the OpenCode source off from settings at any time.
Questions, answered
Can UsageAtlas track OpenCode token usage?
Yes. UsageAtlas reads locally stored OpenCode tokens, requests, models, and sessions, then shows them in the same dashboard as your other supported tools.
Does UsageAtlas show OpenCode cost?
When OpenCode does not report a cost, UsageAtlas estimates one from token counts and bundled pricing and labels it as an estimate. Confirm billing with the relevant provider.
Is OpenCode data sent to UsageAtlas servers?
No. OpenCode activity is processed on your device. UsageAtlas does not require an account, and prompts and responses are not read.