← Weathered Pier LLC
Small, independent practices — and stress-tested against a hospital that never closes
InVets is focused on the independent small-animal practice — typically one to four veterinarians, and comfortably up to eight: an owner who still sees patients, a front desk handling the phones and the counter at once, and technicians moving between exam rooms all day. These are the clinics the large software companies are not built around.
The hardest case we design against is the 24-hour emergency hospital. It has no quiet Sunday to take an update in, no overnight window, and its worst hour — one veterinarian on duty at three in the morning — is exactly when a system going dark costs the most.
Four pillars, and each one is a consequence of how InVets is built
Modern, fast, and few clicks
InVets looks and feels like modern software — clean, calm and quick to read, a long way from the grey screens many clinics still work in. Every screen reads from data already on the machine, so moving between pages is instant, and everyday tasks take a fraction of the clicks.
HOW → §3AThe building keeps working when the connection does not
Every change is saved on the workstation the moment it is made, then sent to the cloud in the background. Check-in, charting and check-out carry on through an outage, and the few things that must wait for the connection say so plainly.
HOW → §3AUpdates that never ask the clinic to stop
The software updates while it is being used, without any downtime. No installer, no workstation visit, no maintenance window.
HOW → §3DTransparency
Every price is public, every change to the record is visible, and our status page reports what actually happened. Nothing about the product depends on a clinic not looking closely.
HOW → §3COffline works the same way as online — with the same rules
The usual way to build offline software is to keep two copies of the database and reconcile them afterwards, deciding whose version wins when they disagree. That is where colleagues’ work quietly disappears, and where records drift apart in ways nobody can explain.
InVets does not reconcile copies. Each change is checked by the InVets service on the workstation and saved there first, and a background process sends it to the cloud whenever the clinic is online. The rules are the same online and offline, and they are enforced in the building.
Online or offline, the InVets services check every booking against every desk in the building at the moment it is made. If two workstations try to book the same surgery room, the second is stopped at the desk — told straight away who holds the room, and picks another. It needs no internet connection.
A small InVets service runs on each machine in the building. The services find each other on the clinic’s network: they keep the office signed in together, settle which desk holds each room and appointment slot, and each one holds the work waiting to be sent. The browser is only the window: closing it, or having it crash, loses nothing. There is no server in a closet and nothing anybody has to switch on.
The right AI for each job — not one model for everything
We don’t use a one-size-fits-all AI. InVets uses different AI models, at different costs, for different jobs within the software — to produce the best possible result at the lowest cost. Some of that work runs on the workstation itself and some in the cloud, and testing decides which.
Arithmetic is arithmetic. Dose calculations and every other calculation in InVets are plain mathematics, worked out the same way every time — not AI.
A visit recorded offline is never lost
If the connection drops mid-visit, the recording is saved on the workstation and held. When the clinic is back online it goes to the cloud, is transcribed, and the notes are added to the record automatically; the workstation’s copy is then removed. The audio is kept for thirty days and then deleted, under a published retention policy — the full raw transcript is kept permanently, as part of the record.
A bulletproof audit trail
Every change to every piece of data is saved, and becomes part of the record’s history. The first value entered, the AI-polished version a vet chose to keep, and the colleague who later corrected the wording are three separate entries — not one field overwritten twice.
Money follows ordinary accounting practice. A posted invoice or a taken payment is corrected by a further entry, never by editing the first.
No maintenance windows. No downtime for an update
InVets is built to run on two servers at all times, and an update replaces them one at a time, each checked healthy before the next is touched — so the clinic is never without a working system. This is established, best-practice engineering, and it is our stated deploy policy: updates land without anyone having to stop.
Many clinics still run software on a server in a back room, which cannot do this at all. The cloud products in this market carry repeated complaints of outages that follow their own releases, and of downtime during server updates.
Our outage should not be their outage.
Which is why we will not print an uptime percentage we have not earned. A vendor in this market published a status page reading 100% through months its own users described repeated multi-hour outages, and the market has learned to discount the number. What we offer instead is structural: if our cloud has a problem, the building keeps working from the workstations while we fix it — backed by a clearly defined service level agreement, so the commitment is in writing.
A status page that tells the truth
InVets keeps a fully transparent status page — maintained without exception and accurate, showing every real issue, past and present, including the ones no clinic noticed. When something looks wrong, a clinic can check it and believe what it says.
The mechanism is what makes it believable. The page is driven by an automated check from two regions, on infrastructure separate from the app, and an incident opens itself. A person writes the explanation afterwards — never the decision to admit it.
The failure condition is written down too: an outage that an outside monitor sees before our own page does. A status page that reports on a measurement, not a judgement call.
Three kinds of incumbent, three different weaknesses
Lightning strikes nearby, the power flickers back on and the internet is gone, and there are sick animals in the building. With InVets, the clinic carries on: check-in, vitals, notes, charges, walk-ins, check-out and cash payments all keep working. Only the few things that genuinely need the connection — card payments, controlled-substance entries — wait for it. When the connection returns, InVets quietly catches everything up.
Downtime around updates is the unserved gap — and it is measurable
Third-party status monitoring and published review aggregation — not our own measurement.
Transparent, top to bottom
Every price public. The card processing rate stated as a number. What is included listed in full — including what is not, such as client texting, which is optional and charged at carrier cost. AI transcription and AI page summaries are included.
And we are deliberately not the cheap option. A practice is protective of its data and its workflow, and a product priced as a bargain invites the question of what was left out — security, stability, somebody to call.
Transparency is the differentiator; being cheapest is not the goal.
What InVets is made of
InVets is in active development at Weathered Pier LLC, ahead of its first clinic pilot. It is built as one platform, end to end, on the architecture this document describes.
The workstation-first save path and its ordering guarantees · the on-machine working copy · live updates across every screen · the clinical, scheduling and financial model.
Each practice’s data kept separate, enforced in the database · a tamper-proof audit trail on every change.
The AI pipeline, and the system that chooses the right model for each job · offline-capable staff and client apps · rolling updates with no downtime.