True Cost to Phone

What an iPhone 18 Pro really costs over 36 months — plan, phone, one-time fees, and the trade-in you hand over — compared across Tello, Mint Mobile, T-Mobile, AT&T, Verizon and Xfinity Mobile.

A static tool in the Rising Company Daylight theme. No build step: index.html + app.js + calc.js read data/pricing.json.

The formula

total = monthly plan × 36
      + phone price − promo credits        (buying from Apple: − trade-in value instead)
      + one-time fees                       (activation · device connection charge)
      + trade-in given up                   (maker's trade-in estimate of the phone the carrier or Apple takes)

The phone you trade in is out-of-pocket at what Apple would pay for it, whichever route takes it. Handed to a carrier for bill credits, an iPhone 16 Pro costs you the $510 Apple offers. Traded to Apple, the $510 comes off the new phone and is counted back as the phone given up — so Tello unlimited with an iPhone 18 Pro Max is $25 × 36 + ($1,299 − $510) + $510.

Every (carrier, plan, route) combination is priced: the "buy from Apple with trade-in, bring your own" route on every plan, plus each promotion the plan qualifies for. Sorted cheapest first.

Lines and phones

Lines (1–5) prices the plan as an account total for that many lines. Each line picks its own new phone (or none) and its own trade-in (or none). Promotions credit each line on its own trade-in; a line without a qualifying trade-in pays full price on that route. One-time fees are per line regardless. T-Mobile promotions credit at most four devices per account. Xfinity's first-year price covers one line; Mint's $10 intro 6 months applies per line. Verizon's myPlan gives new 3+-line accounts a $10/mo account credit for 36 months; Simplicity is one flat price per line.

The Extra savings card holds account-level toggles:

• Switching carriers — new line + port-in, which unlocks port-in deals and intro prices.
• Costco member — buy through Costco. T-Mobile: $150 Shop Card + $250 prepaid Visa per financed phone when switching to Experience Beyond with a port-in (Visa max 4), $75 Shop Card per new line without a financed phone, $100 Shop Card per financed phone when upgrading an existing line, device connection charge waived. AT&T: $250 bill credits + $100 Shop Card per ported line with a financed phone (replaces the $200 online credit), activation waived; bring-your-own does not qualify. Cards are counted at face value. Costco sells no Verizon, Xfinity, Tello or Mint plans.

An Xfinity trade-in credit field appears on a line whose trade-in Xfinity has not priced. The route list can be filtered to one carrier; the filter shows every route for that carrier with its overall rank.

Three tabs

• Routes — every (carrier, plan, route) combination priced for your situation, cheapest first, top three per carrier until you ask for all.
• Compare — up to five routes picked with the Compare button on any row, side by side (total, per month, day one, plan, phones, credits, Apple trade-in, fees, trade-in into the deal, Costco, requirements, lock status) plus a month-by-month out-of-pocket chart: buying from Apple pays the phone on day one, a carrier spreads it over the term. Picks are in the URL as cmp=.
• Plans — pick up to five plans across carriers and see them side by side (monthly, intro price, cost over the term, high-speed data, talk, fees, how the phone is bought, how many phone deals the plan qualifies for right now) plus a cumulative plan-cost chart that makes intro pricing visible. Colors follow pick order. The pick is in the URL.

Run it

python3 -m http.server 8000      # then open http://localhost:8000
node --test tests/               # calculator + analytics tests

The page fetches data/pricing.json, so it needs to be served — file:// blocks fetch.

Data

data/pricing.json was captured from the carriers' own pages on 2026-09-21 with agent-browser (the carrier sites block plain HTTP fetches; a headed session gets through). Sources are listed in meta.sources and on the page.

What's in it:

• phones — iPhone 18 Pro and iPhone 18 Pro Max storage tiers with one retail price each. Apple, T-Mobile, AT&T and Verizon list the same figure (carriers show it as $x.99); Xfinity does not publish one, so Apple's is used everywhere.
• tradeIns — the phone you hand over, valued by its maker's own trade-in program ("up to" values, top storage, good condition): Apple Trade In's full iPhone table; Samsung's standalone trade-in (run by Likewize, pays without a purchase) for Galaxy S20–S25, Note20, Z Fold3–7 and Z Flip3–7; the Google Store estimator for Pixel 6–10 (Google pays the same whichever Pixel is bought). Each entry names its brand and valueSource. Galaxy S26 and Pixel 11 are in neither maker's program yet, so they are not offered.
• carriers[].plans — monthly is the account total keyed by line count ("1"…"5", T-Mobile Beyond to "8"). T-Mobile and AT&T totals were read from their plan pages per line count; Xfinity, Tello and Mint charge the same per line. A plan with no price for the requested count is skipped rather than guessed. premiumDataGb and minutes let the calculator filter by usage.
• carriers[].promos — per-plan eligibility, requires (trade-in / new line / port-in / existing line), bill-credit term, and trade-in tiers keyed by device id, with an otherwise tier for anything not listed. stackable promos (AT&T's $200 online credit) add on top of a main promo and, with appliesToByod, on top of a bring-your-own phone.
• Anything a carrier did not publish is flagged, not guessed. Xfinity's trade-in credits were read at checkout on the crawl date for a new Mobile Plus line: every phone in the trade-in table, iPhone 8 Plus through iPhone 17 Pro Max and every Galaxy and Pixel, quoted the same $930. Xfinity quotes upgrade trade-ins only to a signed-in account, so its "up to $500" upgrade credit is verified: false. Xfinity's separate iPhone Duo credit was quoted at checkout for every trade-in too: $1,000 for an iPhone 14 or newer and the newest Galaxy and Pixel flagships, $500–$700 for older phones, and nothing for a Pixel Fold (a one-time $50 instead). Credits only kick in on Mobile Plus.
• promos[].models — a promo that names models credits only those new phones. Most T-Mobile offers name the 18 Pro and Pro Max; T-Mobile's Beyond trade-in and port-in offers, every AT&T and Verizon offer, and Xfinity's no-trade-in and upgrade offers also cover iPhone Duo; Xfinity's trade-in credit for a new line has a separate Duo offer.

Things worth knowing that the marketing copy hides:

• T-Mobile's "up to $1,000 off" needs an iPhone 15 Pro or newer. An iPhone 14 lands in the $730 tier on Beyond 2.0, $630 on Experience More 2.0, $300 on Essentials 2.0.
• AT&T's $1,200 needs a trade-in AT&T values at $230 or more — iPhone 14 Pro Max, 15 or newer — and only on Premium 2.0 or Elite 2.0. An iPhone 14, 14 Plus or 14 Pro gets $930 there; Extra 2.0 tops out at $930, Value 2.0 at $500.
• Xfinity's deals require Xfinity Internet and the Mobile Plus line; Mobile Select is cheaper but unlocks no phone deals. Its $930 trade-in credit is for a new line only — an existing line upgrading gets up to $500 with a trade-in, or $200 without. All end 11/10.
• iPhone Duo starts at $1,999. AT&T, T-Mobile and Verizon carry their 18 Pro trade-in credits over to it ($1,200, $1,000 and $1,000). Xfinity's "up to $1,000" Duo credit drops to $500–$700 for an iPhone 13 or older.
• Verizon's "up to $1,000" is the new-line credit on Unlimited Ultimate for an iPhone 14 or newer (any working condition); upgrading an existing line gets $840. Unlimited Plus pays $840 / $660, Unlimited Welcome $480 / $300. An older phone gets half on a new line and nothing on an upgrade. Through Apple, a new Verizon line can skip the trade-in and still get $840 / $660 / $360 on Ultimate / Plus / Welcome. Simplicity has no phone deals at all — its pitch is the flat $30/line switcher price with the phone bought from Apple. Simplicity Pro ($50/mo on top, phone returned every year) is a lease, so it is not a route.
• Mint Mobile is priced on its 12-month prepaid plans; new customers get any 6-month plan for $10/mo for the first 6 months, then $30/mo for Unlimited on a 12-month plan. 3-month terms cost more.
• Taxes are excluded everywhere. T-Mobile, AT&T and Verizon prices are with AutoPay (AT&T and Verizon also paperless); Xfinity prices include the $10 internet-customer discount. Verizon's $40 activation fee is counted as $0 because the free Loyalty opt-in waives it.

Planned

The calculator already takes these as inputs; the page just pins them for now.

• More phone models beyond the 18 Pro, Pro Max and Duo (phones[]).
• Choose the trade-in device — present on the page today, backed by Apple's full table.
• Monthly data and minutes needed (minDataGb, minutes) so Tello's smaller plans and AT&T Value 2.0 come into play.
• T-Mobile Essentials 2.0 at 2 lines and Essentials Saver beyond 2 lines are not published on the plan page (4 and 5 lines are priced from its "Essentials 4 Line Offer").

The social card

Every input is in the URL, so a shared link opens on the situation the sender priced. The link preview has to carry its weight: og.png is a picture of the page's own carrier chart — same segments, same colors, same cheapest-first ranking — so the point survives a timeline and the click is verification rather than curiosity.

og.js holds the card's data model (CARD_SCENARIO, cardModel). The scenario is the one the page itself computes on first load: one line, an iPhone 18 Pro 256, an iPhone 16 Pro to trade, not switching. A card that disagrees with the page it links to is a worse card.

Regenerate it whenever data/pricing.json moves:

python3 -m http.server 8000
agent-browser set viewport 1240 720 2          # 2x — the card is 1200x630 CSS px
agent-browser open http://localhost:8000/tools/og.html
agent-browser wait --load networkidle
agent-browser screenshot "#card" og.png        # writes 2400x1260

Then restamp the landing page, which owns eight strings derived from the data — the ?v= cache stamps, the headline spread in the tags, the six totals in the alt text, the two dataset dates, the no-JS hero line and the footer:

node tools/restamp.mjs           # rewrite in place
node tools/restamp.mjs --check   # exit 1 if index.html has drifted

Scrapers cache og:image by URL, so without a new stamp Facebook, Slack and X keep serving the old numbers for days — the exact failure a weekly re-crawl exists to avoid.

The whole refresh is written up as a skill: .claude/skills/refresh-carrier-pricing.

tests/og.test.mjs pins all of it: the scenario, the arithmetic behind each bar, the numbers printed on the committed card, the stamp against meta.crawledAt, and the spread quoted in the tags. Moving the data without regenerating the card is a failing test, not a stale share.

Generated pages

The tool answers every situation at one URL, but the question arrives in several hundred shapes — "verizon vs t-mobile iphone 18 pro", "is xfinity mobile worth it". seo.js builds a model for each of those, and tools/gen-seo.mjs renders them:

node tools/gen-seo.mjs     # 15 compare/ pages, 6 carrier/ pages, sitemap.xml, robots.txt

Output is committed, like og.png. Regenerate it whenever data/pricing.json moves; tests/seo.test.mjs fails if the committed pages carry a stale crawl date.

Each page states numbers the calculator produced, names its sources, and deep-links into the tool pre-loaded with the situation it describes. The carrier pages carry the part the carriers' own pages never show: each promotion priced against buying the same phone outright on the same plan, which holds the plan constant so the difference is the deal and nothing else.

Two scenarios drive everything, both in scenario.js:

• HEADLINE_SCENARIO — one line, an iPhone 18 Pro 256, an iPhone 16 Pro to trade, not switching. What the landing page computes on first load, so the card agrees with it.
• NO_TRADE_IN_SCENARIO — the same line with nothing to trade. The generated pages lead with it, and their tool links carry it (l=iphone-18-pro-256:-) so the tool opens on the numbers the page quoted.
• SWITCHER_SCENARIO — the same line switching in with an iPhone 16 Pro to trade. Every promotion requires a port-in, a trade-in or both, so the no-trade-in scenario filters all of them out; a page about whether a deal is worth taking has to price the case where it is on the table.

Both spread in USAGE_FLOOR (50 GB premium data, unlimited minutes) — the floor app.js has always applied. It lives in scenario.js because leaving it out prices plans the page never shows: the social card once shipped a Tello plan with 300 minutes and no data as the cheapest route.

Measurement

PostHog, provisioned through the Vercel Marketplace (vercel integration add posthog, US region). analytics.js holds the whole of it.

The project token in analytics.js is a public write-only key — the same one PostHog's own snippet puts in the page source. There is no build step to inject it from the environment, so it is written in. The Marketplace also set *_POSTHOG_PROJECT_TOKEN and *_POSTHOG_HOST on the Vercel project for anything that later needs them.

Nobody signs in here, so nobody is identified: person_profiles: "identified_only" keeps readers anonymous. shouldTrack() drops localhost and .local, so serving the folder locally never reaches the numbers.

Alongside PostHog's automatic pageviews and click autocapture:

• situation_priced — the situation someone actually priced (lines, phones, trade-ins, switching, Costco, carrier filter) and what won it (carrier, plan, route, total, spread). Debounced, and a situation already reported is not reported again until something changes — tab switches re-render without being a new data point.
• tab_selected · route_compared / route_uncompared · carrier_filtered · routes_expanded · compare_tray_used · summary_card_jumped

Counts and ids only. No prices typed by the reader, no free text, no identifiers.

Contributing

Everything the page shows comes from data/pricing.json, and each entry names its source. If a price has moved, a promotion is missing, your carrier is not here, or a checkout showed you a trade-in credit the carrier does not publish, open a pull request or an issue at https://github.com/rising-company/true-cost-to-phone. Keep meta.crawledAt honest, note where a number came from, and run node --test tests/ — the tests pin the arithmetic and a few facts worth not losing (T-Mobile's tiers, Xfinity's one-line intro, the Costco rules).

Tests

node --test tests/ — zero-dependency, matching the other static tools in this workspace (www/stream). This deviates from handbook ADR-0006 (Vitest) because there is no package.json to hang it on; the layout (tests/ folder, node environment) is the same.