feature: answer the question the README already promises — "is my plan worth it?"

[pitimon]

README:32 already makes this claim:

Subscriptions hide the number entirely. A flat monthly plan shows you a quota bar, not what your usage would have cost. TokenTracker prices every token against public model rates, so you can see whether the plan is a bargain or a subsidy you have outgrown.

The product does not compute it. It shows what the usage would cost at list price, and it shows a quota bar, but the comparison — the actual decision — is left as mental arithmetic against a plan price the product never asked for.

Every input already exists: per-source cost per window, and the pricing tier telling you how much to trust it.

Proposal

A per-provider plan-value card:

• Plan price entered by the user in Settings, stored locally like everything else — never sent anywhere, no plan catalogue to maintain and go stale.
• Shows: Claude Pro — $20/mo · your usage this month would cost $47 at API list price.
• Rolls up to one headline across providers.

The labelling is the hard part, not the arithmetic

This number is easy to over-claim and must not read as "you saved $27":

• It is list-price-equivalent, not a counterfactual bill. Anyone actually on the API would use it differently.
• It must inherit the pricing caveats already shipped in feat(pricing): in-process refresh + report how each price resolved #92: if any model in the window is unpriced or fuzzy-matched, say so on the card. unpriced_models and fuzzy_priced_models exist for exactly this.
• Under-usage is as useful as over-usage — "$3 of usage on a $20 plan" is a downgrade signal — so the card must read neutrally in both directions, not celebrate one.

If the honest framing cannot be made to fit, the right outcome is to soften README:32 rather than ship a number that flatters.

Why this rather than alerts

Considered and rejected: a budget-alert daemon. Notifications are three-surface work (web, Swift, C#), the acute case is already covered by the quota chips, and the value here is reflective — a number you look at when deciding whether to keep paying, not a ping.

Priority

Lowest of the current set. #100/#101 close a live privacy defect and its guard; #102/#103/#104 pay down real debt. This one adds surface. It is on the list because the README already sold it, which makes it a truth gap rather than a nice-to-have — but it should be built after the debt, not before.