[roguetrader777]

Summary

OpenQuota's usage view was limited to preset periods and separate provider cards. It could not show total usage across every enabled provider for an arbitrary local-time range, usage since a real weekly or observed banked reset, measured task averages, or a clear comparison with two premium seats plus entered API spend.

Additions

• Add a Custom period beside Today, Yesterday, and 30 Days, using the existing Total Spend ring and provider legend.
• Select all enabled providers by default, while still allowing one provider to be inspected.
• Add arbitrary local date/time ranges using an IANA time zone.
• Add since last reset views based on live provider reset data, including an observed Codex banked-reset marker after a successful reset claim.
• Add measured average cost and duration per completed Codex or Claude task from content-free local event metadata.
• Add local reset/update timestamps, range provenance, incomplete-data diagnostics, and comparisons for two current premium seats plus entered API usage.

Fixes and safeguards

• Prevent Today from remaining visually selected when Custom is active.
• Resolve DST gaps and overlaps deterministically instead of silently shifting the requested range.
• Use the live weekly-reset boundary for daily-only providers instead of unrelated Custom form values.
• Fail explicitly when provider data is unavailable or only daily precision is supported; hourly detail and task averages are never invented.
• Keep raw prompts and responses out of both the frontend response and analytics cache.
• Reuse the existing provider discovery, storage, pricing, and Tauri command layers; no dependency was added.

Verification

Tested on Windows 11 x64 through WSL with the Windows Rust toolchain:

• corepack pnpm verify:versions — PASS
• corepack pnpm verify:contracts — PASS
• corepack pnpm lint — PASS
• corepack pnpm check — PASS (0 errors, 0 warnings)
• frontend test suite — PASS (240/240), followed by the final affected UsageAnalytics regression suite — PASS (7/7)
• corepack pnpm build — PASS
• cargo fmt --manifest-path src-tauri/Cargo.toml --all -- --check — PASS
• cargo clippy --manifest-path src-tauri/Cargo.toml --all-targets -- -D warnings — PASS
• cargo test --manifest-path src-tauri/Cargo.toml --all-targets — PASS (550/550)

corepack pnpm verify reached the frontend Vitest phase, where the default unbounded worker launch stalled in the WSL/NTFS environment. The same required checks passed separately with bounded workers. Upstream GitHub Actions is awaiting the maintainer's first-time fork approval; no failed CI job exists on the current HEAD.

Reference prices are current published USD prices and are shown only as local estimates, not provider billing data:

• OpenAI Business pricing
• Cursor Teams pricing
• Anthropic pricing

Screenshots

Full dashboard with Custom applied, matching the existing main screenshot format (synthetic example data):

Custom range configuration dialog:

Checklist

• This pull request is focused on one change.
• I added or updated tests where behavior changed.
• I ran the relevant checks, including corepack pnpm verify when possible.
• I did not include credentials, personal data, build output, or unrelated changes.