Ingestion
A continuous read of public reporting communities, polled every 30 seconds so a delivery batch never sits waiting. Every message is recorded before anything is done with it, so nothing is lost if a later stage fails.
Home / How it works
How it works
No magic, and nothing that touches your visa account. Here is the whole chain, including the parts of it we do not control.
Thousands of people check the official scheduling site every day and post what they find — screenshots of open slots, and just as often, that there is nothing. We read that stream continuously, around the clock.
Positive sightings almost always arrive as screenshots. Our extraction reads the consulate, visa type, date, time and — where the report shows one — the number of seats, and turns it into structured data. What it cannot read, it marks as unread rather than guessing.
Consulate, visa type and date range, plus the seat count where the report stated one. Around a third of sightings never state it; those still reach you, labelled as unstated, because an absent number is not evidence a slot is too small.
No app to open, no extension, no inbox to check. The alert arrives where you already are, stamped with how long ago it was seen and how many people confirmed it.
What an alert says
A report is whatever an applicant happened to type or screenshot, so some fields are simply missing. The rule we hold to is that a gap is shown as a gap. An alert never carries a number we did not read.
| Field | Where it comes from | If the report never said |
|---|---|---|
| Consulate and visa type | Stated in the report, or read from the screenshot. | The report is unusable and never becomes an alert. |
| Date and time | Read from the calendar in the screenshot. | The alert says “date not stated in the report”. |
| Seats | Read from the screenshot when the calendar shows a count. Around two thirds of sightings state one. | The alert says “seats visible: not stated in the report” — and still reaches you, because an absent number is not evidence a slot is too small. |
| Age of the sighting | The timestamp on the message we read it from. | Always known. It is the one field that cannot be missing. |
| Corroboration | How many independent reports named the same slot. | Shown only when more than one did. |
Reporting a slot
The fastest sighting we ever get is one nobody had to post to a group first. Send us a screenshot of the calendar — that is the whole report. We read the consulate, the date and the seats off the picture, show you what we read, and on your say-so it reaches everyone it matches within seconds.
| If you are | Each report earns | Capped at |
|---|---|---|
| On a paid plan, inside it | 1 free day on your subscription | 7 days a month |
| Free, on trial, or lapsed | 2 extra alerts | 10 alerts a month |
We ask a question only when the picture cannot answer one. Usually that is none. The exception is the visa type: the booking site makes you choose it before it shows you a calendar, so a screenshot of one never displays it — and if you are registered for a single visa type, we already know. A slot somebody already reported earns nothing, and neither does a screenshot we have seen before.
Under the hood
Built so that a failure in one part cannot stop an alert reaching you from another. Delivery outranks everything else in the system.
A continuous read of public reporting communities, polled every 30 seconds so a delivery batch never sits waiting. Every message is recorded before anything is done with it, so nothing is lost if a later stage fails.
Text reports and screenshots both go through structured extraction. One message commonly yields several facts — a single 'nothing at any consulate' report is ten separate observations once it is broken down by consulate and visa type.
The same slot gets reported by several people within seconds. We collapse those into one event and count the independent confirmations, which is what lets an alert say how well corroborated it is instead of just asserting it.
Subscribers report slots they find themselves, with a photo of the calendar. These are the fastest sightings we get — nobody had to post them to a group first — and they are the one source we can ask questions of.
Your consulates, your visa types, your date window — and your party size against the seat count, on the sightings that state one. Only what survives reaches you.
A WhatsApp message, stamped with when the slot was seen. No app, no extension, no inbox.
The service watches its own sources. If a source stops delivering, we know within hours — because a quiet market still produces messages, and true silence means something is broken rather than that nothing is available.
Being straight with you
Our reports come from applicants, so there is a gap between a slot appearing and somebody noticing it — usually under a minute. A bot polling the booking site directly would beat us over that gap. It would also be breaking the scheduling system's terms, and one enforcement action from disappearing.
Nobody can. Slots in this market can be gone in seconds. What we can do is tell you exactly when it was seen and how many people confirmed it, so you decide whether to race for it. A service that implies otherwise is selling you certainty that does not exist.
We are live for H-1B and H-4 in-person interviews because that is where reporting is dense enough to be genuinely useful. We would rather say so than take money for a visa type we cannot properly watch.
Questions about the mechanism
From applicants. Thousands of people check the official scheduling site every day and report what they find to public communities — usually as a screenshot when something is open, and as a short text note when nothing is. We read that stream continuously and turn it into structured records.
We do not poll the booking system, and we hold nobody's scheduling credentials.
It would be, by perhaps a minute. It would also be automated interaction with a system whose terms forbid it, which is why services built that way spend their lives fighting anti-bot measures and can disappear overnight.
There is a second, less obvious reason. A poller sees one account's view of the calendar. A crowd is thousands of people checking from different accounts at different moments, so between them they notice openings no single watcher would. Breadth partly compensates for the speed we give up.
Positive sightings almost always arrive as images, so an alert service that only reads text misses the events that matter. We run vision extraction over every image: consulate, visa type, date, time and, where the calendar shows it, the number of seats. Anything the image does not show is recorded as unread rather than guessed, and the alert repeats that to you.
Yes, and it is the fastest report there is — no waiting for somebody else to post it. Send us a screenshot of the calendar; we read the consulate, the date and the number of seats straight off the picture. It goes out to everyone it matches within seconds.
Usually that is the entire conversation. We ask a question only when the picture cannot answer one — most often the visa type, because the booking site makes you choose that before it shows you a calendar, so a screenshot of one never displays it. If you are registered for a single visa type we do not even ask that.
You do not need to be on a paid plan. Anyone with alerts switched on can report.
If you are on a paid plan and inside it: 1 free day added to your subscription, up to 7 days a month. Everyone else — free, trial, or a plan that has run out — earns 2 extra alerts, up to 10 a month.
The split is not favouritism. Extra days are worth nothing to someone who is not on a paid plan — the delivery rules would still treat them as free — so paying them in days would be thanking them with a number that changes nothing.
We do not tell you how many subscribers it reached, and that is on purpose rather than an oversight: anyone can sign up free, so a reply naming a number would hand a competitor our subscriber count for that consulate and visa type, one report at a time.
Because the photo is the report. There is nothing to type, so there is nothing to claim without evidence — a screenshot is the only thing you can send, and producing a convincing false one is real work. That is what stands in for the paid-only rule we used to have.
It also has to show something. A picture of an empty calendar is read as no availability and is not sent to anybody, because it is not a slot.
Four things, and they run in order of what they cost us. There is a daily limit per person. The same screenshot sent twice is recognised and refused. A slot somebody already reported earns nothing, because it is corroboration rather than news. Only what survives all three is checked against the photo.
The monthly caps are the backstop: whatever happens, one person cannot earn more than 7 days or 10 alerts in a month.
Anything nobody reported. If a slot opens at 4am and no applicant happens to be looking, we do not know about it. We also cannot tell you a slot is still there — by the time you read an alert, someone may already have taken it. That is why every alert carries the time it was seen rather than implying it is still open.
Tell us the consulates, the visa type, your date window and how many people are on your application. We do the watching.
First 5 alerts free · cancel from WhatsApp · refund guarantee