The desk.

An independent desk for Indian points-and-miles geeks. The subject is travel value — miles, transfer partners, lounges, premium metal. The catalogue is wider: every Indian card we can find an issuer document for, cashback included, because the card in your pocket is usually not the one a travel blog writes about. What this is not is a "10 best cards" listicle, or a lead-generation funnel wearing a blog's clothes.

Zero affiliate links. No display ads. No sponsored placements. Nobody pays for a rank, a verdict, or a kind word — that is a locked, written decision, and the methodology page spells out what happens if it ever changes (disclosure first, rankings untouched). When we say a card is a Cancel, no commercial relationship is whispering otherwise, because there isn't one.

Nobody types a rank either. No card file carries one; a file that tries to fails the build.

The loader does it — sorted on can-you-still-get-the-card, then our editorial order, and numbered 1 to 1119 on every build.

We describe practices here, not personalities. There is no team page, no founder origin story, no stock photo of people pointing at a laptop. The site's credibility is designed to rest on things you can open in another tab — the issuer document behind each fact and the date we last read it, a published methodology, and a dated, sourced tracker of every change — not on who is behind the keyboard.

The news desk is machine-written, and says so. It phrases our own dated tracker events and may state nothing the data does not already contain — five fences, set out on the methodology page. Rankings, valuations and verdicts are curated by hand.

Everything the site shows is generated from a single set of hand-curated, version-controlled data files — card terms, currency valuations, transfer routes, lounge access, 376 dated tracker events. Every page, every rank, every ₹ figure is computed from that data at build time. To change what the site says, you have to change the data — and every change is a reviewable diff.

Every card cites its paperwork. All 1119 of 1119 card pages carry a Sources section — 3800 official issuer documents (T&C, MITC, program pages) plus community corroboration, each with the date we last checked it. A fact we can't verify doesn't get invented; the page renders "not published" instead.

A number we cannot source is blank, never zero. On 404 of those 1119 cards, no document we could find states the annual fee, or states how the card earns.

Those cards keep their page and lose their rank — carried, not ranked, and never fed to the calculator.

Which is what makes a ₹0 here a fact. 294 cards are genuinely free for life, and "free" must never look like "we don't know".

And none of it is allowed to go quiet. The oldest review date on the whole board is 26 July — 66 days ago. Nothing here has gone unread for longer than that, and every card prints its own date.

The data is guarded twice. A build-time audit is a hard gate — no build ships if a transfer ratio fails to parse, a valuation escapes its own floor–ceiling range, or a cross-reference dangles. And a daily data-quality watchdog re-audits what production is actually serving every morning and files a report to the desk the moment anything drifts.

What that means for you, precisely. Data, code, audit scripts and edit history all live in version control, so every number has a dated before-and-after — but the repository is private, so treat that as our audit trail, not your evidence. Yours is on the pages: 4059 source lines across 1119 cards, each naming its document and the date we read it, and 376 dated, sourced entries in the tracker. If a number here traces to neither, don't trust it — tell us.

Which board a card lands on

Two lists, and a card cannot cross. The value model prices what a card earns and cannot tell statement credit from airline miles; one list would rank a cashback card above Infinia.

244 travel cards you can apply for today are ranked here. 153 that pay you back in rupees get their own board and their own #1.

Five gates sort them — a lounge quota, a point that reaches an airline or a hotel, a travel co-brand, a travel accelerator, a forex markup under the Indian norm. Any one is enough.

Every gate asks for evidence FOR, never against. A card we can prove nothing about files as cashback: mis-filing a travel card costs us traffic, mis-filing a cashback card costs you money.

And a transfer route has to beat doing nothing. An exit worth less than the points left alone is a decoy, not a door, and it does not make a card a travel card.

The one number nobody publishes

Every rank is downstream of one figure per currency — what a point is worth in rupees. No issuer publishes it. No document anywhere states it.

A house value is consensus, reasoned from an award chart, a portal rate, a transfer ratio or a currency peg — which makes it the easiest number on a site like this to invent.

So the valuations cite their evidence too. All 159 of our 164 currencies carry a sources block of their own — 327 documents, each dated, shown on the currency's programme hub beside the value they argue for.

The other 5 are rupees at par; the rupee needs no citation.

How the numbers themselves are made — the ₹1-per-mile anchor, the ranking engine, the verdict scale — is disclosed in full on the methodology page.

A wrong fee or an inverted transfer ratio is worse than no site at all, so the policy is simple: when we get something wrong, we fix it where you can see it. A corrected fact ships with a fresh "checked" date against the issuer document it was re-read from. A change in the world gets a dated, sourced entry in the tracker. And when the number we had was materially wrong, it gets a post that names it. We do not quietly edit a number and hope nobody kept a screenshot — and we don't ask you to take a private commit log on faith either. Two examples you can open right now:

• The Atlas closure date. Axis closing the Atlas to new applicants first surfaced as a rumour carrying a "1 September" cutoff, and our story initially carried that future date as if it were the event. QA caught it and the post was re-dated to when the closure was actually confirmed — and it now says out loud which half was rumour and which half Axis confirmed, so the correction is readable rather than asserted: read the post .
• The July 2026 re-verification. We re-checked every card against issuer PDFs and found ~35 factual errors in our own data — including ones where we'd been too harsh: an Amex reward rate understated by half, a "no fee waiver" claim that was false, a transfer route we said didn't exist that does. Every one shipped, and the sweep is legible on the cards themselves: open any card page, scroll to Sources, and each line carries the document and the date it was last read.

When the world changes rather than our data being wrong — an issuer nerfs a card, a program devalues — that isn't a correction, it's an event: it gets a dated, sourced entry in the tracker and the affected pages update the same day.

Spotted something wrong?

Tell us — you will be doing every reader after you a favour. The channel is the report form: four fields — the page, what's wrong, the correct fact, and (ideally) a source. It goes straight to the desk's daily sweep; no account, no email (this domain deliberately sends and receives none). If the issuer's own document backs you, the fix usually ships the same day.