Replace Cloud tiered pricing with the v3 single-plan model

[roncodes]

Replaces the 13 tiered "resource unit" Cloud plans with the v3 single-plan model across the whole site, rebuilds /pricing around it, and replaces the order-volume calculator.

Numbers are taken from the billing implementation, not from the brief's prose: packages/billing/server/src/Support/V3Statement.php and config/plans.php → plans_v3. I checked every figure against those files and they agree.

Important

Do not merge/deploy before the product cutover. plans_v3 currently ships with 'is_visible' => false, with a comment saying Phase 6 flips it in the same release as the console build. So this pricing is not live in the product yet. See Open questions below.

---

The pricing page

Rebuilt around one plan, reusing the existing design system rather than inventing a new one. The "Which one are you?" chooser, self-hosted/professional-services cards, support tiers and commercial-licence callout are structurally unchanged.

Section	What it does now
Hero	The price plainly: $29 a month, plus $5 per driver or vehicle, with the per-delivery point as the supporting line
Billing toggle	Monthly / Annual with the 20% saving shown; one piece of state, rendered both above the plan card and inside the calculator so they can't disagree
Plan card	The four approved bullets verbatim from plans_v3 → highlight_features, plus an "Unlimited, at no cost" chip list
Calculator	Line-by-line estimate — given real prominence directly under the plan
Usage rates	The three meters and their free allowances, shown on the page rather than behind a disclosure
FAQ	Rewritten: do I pay per order, what counts as a driver/vehicle, what if I remove a vehicle, what is a gateway-paid invoice, will I be charged for API usage, monthly vs annual, what happens when the trial ends

Removed: all 13 tiers, "Most Popular", "Show all 13 plans", the overage and top-up pack tables, and the "How resource units work" disclosure.

Deliberately not built: the competitor comparison table. The brief made it optional and required date-stamped, signed-off, publicly published competitor pricing. I had no way to verify current Onfleet/Tookan/Route4Me pricing, and inventing numbers is exactly the published-wrong-number risk the brief warns about. The existing per-competitor comparison pages already carry their own competitor figures — see Open questions.

How the calculator works

src/lib/pricing/v3.ts is a direct port of V3Statement.php — meteredAmount(), licensedAmount() and build(), with the plans_v3 rate table. All arithmetic is in cents, formatted only for display.

resources           = drivers + vehicles
billableStorefronts = max(0, storefronts - 1)      // first storefront free

monthly: base 2900,  resource 500,  storefront 500
annual:  base 27840, resource 4800, storefront 4800    // 20% off recurring

ledger  = invoices     * 10                        // every invoice, no allowance
api     = floor(apiCalls     / 100000) * 25        // round DOWN, first block free
webhook = floor(webhookSends / 100000) * 25        // round DOWN, first block free

• Inputs: drivers (5), vehicles (3), storefronts (0), billing interval. The three usage inputs sit behind a collapsed "Estimate usage fees" disclosure — the headline estimate is correct and complete without them.
• Orders input: included and priced at $0.00 — orders are always free, however many you run. It demonstrates the Onfleet contrast far better than a sentence does.
• Annual: usage meters are fed annual volumes (monthly input × 12) because meters are period-to-date; the rates themselves are identical on both intervals.
• Output: a statement, not a number — each line with quantity, unit rate and amount, then a recurring subtotal, a usage subtotal and the estimated total. Zero-value usage lines still render so the reader can see the meter exists and is costing nothing. The storefront line appears only when storefronts > 0 and reads 3 storefronts, first included · 2 × $5.00.

Verified

V3StatementTest.php is ported case for case into src/lib/pricing/v3.test.ts, plus assertions pinning the claims the page makes. npm test (node's built-in runner, no new dependencies) — 12 passing.

I also drove the deployed calculator in a browser rather than trusting the unit tests alone:

Case	Expected	Live UI
25 drivers, 5 vehicles, 2 storefronts, 6 invoices, monthly	recurring $184.00 · usage $0.60 · $184.60	✅ exact
99,999 API calls	$0.00	✅
100,000 API calls	$0.25	✅
1 storefront	adds $0.00	✅ (line shown, labelled)
Annual, same inputs	exactly 20% off recurring	✅ $1,718.40 vs $179.00 × 12 × 0.8

Everything else touched

Pricing model copy

• compare/ (index, vs-onfleet, vs-tookan, vs-route4me) — resource-unit framing → the per-delivery contrast; vs-onfleet's Fleetbase pricing card $25/100 units → $29 + $5; three stale $50/mo "Pricing Model" rows (the old Lite tier) corrected
• platform/fleetops (page + content + JSON-LD Offer), platform/storefront (commission FAQ), solutions/trucking, solutions/roles/executives, partners, company/investors
• contact/sales, developers — trial "capped at 50 resource units" removed; API FAQ now states the real 100k free allowance and $0.25/100k rate
• components/sections/deployment-options — "From $25/mo" → "$29/mo + $5 per driver or vehicle"
• content/docs/community/faq.mdx — Cloud pricing answer and the self-hosted/Cloud comparison table

Now-false claims removed. "No per-seat pricing" is still true (users are unlimited and free) and was left alone, but "no per-driver fees" is now false — drivers are $5. Corrected to "never a per-delivery fee" on the homepage metadata/OG, sections/hero, sections/seo-content, and the Fleet-Ops page.

SEO / structured data

• seo/json-ld.tsx — SoftwareApplicationSchema default price 25 → 29, with an Offer.description covering the per-resource lines so a single number isn't presented as the whole price. No page overrode the default, so every stale Offer on the site moved with it.
• /pricing metadata, OG and Twitter descriptions rewritten; OG image subtitle regenerated. Added an explicit SoftwareApplicationSchema to /pricing.
• URL-encoded OG subtitles (no%20per-driver%20fees) were carrying the old claim past plain-text greps — fixed.

Legal — needs review. Terms §1 and §6 defined and priced Resource Units, so leaving them would have published a contract for a billing model that no longer exists. Rewritten factually in terms of Billable Resources and Metered Usage, keeping the existing structure and register; §5's "20% discount (equivalent to two months free)" corrected to ~ten months' charges for twelve months. Privacy's "Resource Unit consumption" → metered usage and billable resource counts. I am not a lawyer — please have these reviewed.

Fleet-Ops rename. 28 display strings in src/, 88 in content/docs/. Code is untouched: fenced code blocks, inline code spans, Fleetbase\FleetOps\… namespaces, the fleetbase/fleetops repo slug, getModelPackage('FleetOps') and component identifiers all stay. Applied with a script that skips code contexts, then hand-checked the 32 remaining occurrences — all genuinely code.

Verification

• ✅ npm test — 12/12
• ✅ No resource unit / overage / top-up / Micro / Starter / Scale remains as Cloud plan pricing (remaining "overage" hits are the substring in "coverage"; $200/mo hits are the installation Maintenance Plan, a separate product, left alone)
• ✅ tsc --noEmit — 0 errors in changed files. 48 pre-existing errors remain, all PageData in the docs/[[...slug]] route templates
• ✅ Renders correctly at 390 / 834 / 1440. Confirmed via DOM measurement that the page does not scroll horizontally at 375px; the usage-rates table scrolls inside its own overflow-x-auto container by design
• ✅ Signup links unchanged and still resolve to console.fleetbase.io/onboard
• ⚠️ Lint: 83 files fail — but main fails on the same 83. Diffed both: this branch introduces zero new violations. Mostly simple-import-sort and unescaped entities; worth a separate autofix PR
• ⚠️ Build: compiles successfully, then fails collecting page data for /blog/[slug] — Ghost Content API request failed with status 402. I built main in a clean worktree and it fails identically, so this is a pre-existing expired/unpaid Ghost credential, not this change

Two unrelated things in the working tree were left uncommitted: the modified vendor/postman submodule pointer, and an untracked pnpm-workspace.yaml. That file makes every pnpm run script fail with ERROR packages field missing or empty, which is why verification here used npm.

Open questions — please confirm before this ships

1. Launch timing. plans_v3 is is_visible => false pending the Phase 6 console build. What is the cutover date, and should this PR be held or merged-but-not-deployed?
2. Existing customers. Nothing on the page says current customers are already on this model, but nothing addresses them either. Do you want a "what this means for existing customers" note, and what does migration actually look like?
3. Self-hosted ($2,500 one-time) and the four support tiers are untouched, as instructed. They sit below the single plan and read fine — please confirm that's still correct.
4. Competitor numbers. I built no new comparison. But the existing pages still carry unverified, undated competitor pricing — Onfleet "$550/mo Launch plan, 2,000 tasks", plus "Per-driver"/"Per-agent"/"Per-route" model claims for Onfleet, Bringg, Spoke, Tookan, Detrack and Route4Me. Those now sit next to a freshly-updated Fleetbase column, which makes them more load-bearing than before. They need sign-off and date-stamping, or removal.
5. Trial length. The site says 7 days, matching the billing.trial-duration default. That setting is operator-editable and plans.php explicitly warns against copying it into plan config — so this figure can go stale. Worth confirming, or softening to "free trial" with no number.
6. Pallet. Per the accuracy guardrails I kept Pallet out of the pricing page's included-modules line and off the vs-onfleet module list. It is still named as a shipped module on /platform/pallet, solutions/roles/warehouse-managers and a vs-route4me row ("Pallet WMS included"). Rewriting those is well outside this change — flagging it since Pallet is not released.
7. Naming. content/docs/contributing/translations/glossary.mdx previously declared the opposite convention — that FleetOps is canonical and "hyphenated Fleet-Ops may appear in older copy". I flipped it to match the brief and noted that code keeps the unhyphenated form. Please confirm that's the intended direction.

Screenshots

Before/after of /pricing, captured full-page via CDP at identical device metrics from main and this branch.

(Attaching below — desktop 1440px, tablet 834px, mobile 390px.)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
fleetbase-io	Ready	Preview	Aug 8, 2026 9:00am