dsh-token-usage

When you manage several AI provider accounts at once, getting a full picture of your remaining quota means hopping between ChatGPT settings, Grok dashboards, and Ollama Cloud pages. dsh-token-usage consolidates those scattered data points into the DSH Web GUI, so you can view provider-reported billing cycles, percentage allowances, and reset times side-by-side with locally recorded request counts, project attribution, and valuations.

The plugin works in three layers. Provider connectivity handles OAuth for ChatGPT and Grok, multi-account activation with auto-failover for Antigravity, and API credential registration for GLM and Ollama, all over loopback channels with secrets confined to owner-only files; it never reads a browser profile and rejects redirects that could carry credentials across origins. The local usage ledger is built on SQLite and preserves request folding, project attribution, valuation, import and export, backup, and retention policies, while clearly labelling itself as DSH-observed usage rather than a provider invoice. The privacy boundary is strict: no telemetry, no prompt or tool-argument storage, no network requests during ordinary ledger reads, and explicit user opt-in for price updates and quota refreshes.

It suits technical users running three or more model providers inside DSH who want to cross-reference remaining quota against actual daily usage in one table. Workflows that are privacy-sensitive and reject any prompt persistence or third-party telemetry will find a particularly good fit.