fix(usage): distinguish unpriced usage from zero cost

[caezium]

What Changed

• Render fully unpriced usage as — and mixed priced/unpriced usage as a lower bound (≥$X) instead of an exact $0.00 or $X.
• Surface unpriced coverage consistently across the web/desktop and mobile headline, provider, model, chart, and day breakdowns while keeping token totals complete.
• Share the cost presentation rules between clients and add focused all-priced, mixed, and fully-unpriced regression coverage.

Why

The aggregation layer already knows when a model has no matching API rate, but the UI formatted that missing price as zero. That made unknown cost look like a genuine free result and made mixed totals look exact. This keeps the existing API-rate estimate model while making its coverage honest.

Fixes #5799.

Focused verification:

• 33 shared/web/mobile usage tests
• targeted format and lint checks
• @t3tools/web and @t3tools/mobile typechecks
• independent standards/spec review

UI Changes

All screenshots use isolated synthetic transcripts and rates.

Web / desktop

Before	After

Mobile summary and providers

Before	After

Mobile model breakdown

Before	After

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for any UI changes
• No animation or interaction changed, so a video is not applicable

Note

Distinguish unpriced usage from zero cost across web and mobile usage views

• Adds unpricedShare to provider, model, day, and per-provider-day totals in usageMerge.ts, tracking the fraction of records with no matching API rate.
• Adds formatCostEstimate in usageFormat.ts to render — for fully unpriced, ≥$ for partial, and plain USD for fully priced costs.
• Updates web (UsagePage.tsx, UsageProviderChart.tsx) and mobile (UsageRouteScreen.tsx) to use coverage-aware formatting throughout headlines, chart tooltips, provider rows, and breakdown tables.
• Adds guidance text (e.g. ≥ and — explanations) inline when any unpriced records are present.
• Behavioral Change: any cost previously shown as $0.00 for unpriced models or providers now displays — or ≥$X.XX depending on pricing coverage.

^{Macroscope summarized 9920f0e.}

---

Note

Low Risk
Presentation-only changes to usage cost formatting with shared helpers and broad test coverage; no auth, billing, or aggregation logic beyond exposing existing unpriced record counts.

Overview
Usage cost UI no longer treats missing API rates as $0.00 or exact totals when coverage is incomplete.

Shared: formatCostEstimate returns — when everything is unpriced, ≥$… when only part of usage has rates, and unchanged dollar amounts when fully priced. mergeUsage now surfaces unpricedShare on providers, models, daily rows, and per-provider day cells so clients can format honestly.

Web: Headlines, provider rows, model/day breakdowns, and chart hovers use those rules; partial windows get (partial) labels, a short cost guide, and “share of priced cost” where relevant. Missing provider-days show — (no activity), not priced zero.

Mobile: presentUsageCost / presentUsageCostShare mirror the same copy for the chart card, providers, and models, including the cost guide when any usage is unpriced.

Regression tests cover all-priced, mixed, and fully-unpriced cases on shared, web, mobile, and chart helpers.

^{Reviewed by Cursor Bugbot for commit 9920f0e. Bugbot is set up for automated code reviews on this repo. Configure here.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 37ee1a91-8de3-414a-b67b-e92e2b2a93a7

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR modifies billing/usage cost presentation logic, changing how costs are displayed when API rate matching is incomplete. Changes to billing-related display warrant human review to ensure the new presentation (using '—' and '≥' indicators) accurately conveys pricing information to users.

^{You can customize Macroscope's approvability policy. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ddbc7240621222df81c9f9fa64c03a77de65fe38. Configure here.}

Dismissing prior approval to re-evaluate 9920f0e

[juliusmarminge]

Closing as part of the open-PR backlog sweep (wave 1).

Reason: Unpriced-cost UI targets a usage page that has since been rewritten
Related: #9860

Reopen if this is still wanted and you’re willing to rebase onto current main.