Why I'm Building Dentisly: Software That Thinks in Chairs

Build notes on Dentisly, dental clinic software: chair schedules with emergency slots, chairside charts, invoicing with per-line discounts and part payments, five staff roles, and what is not live yet.

Most clinic software is written for general medicine, and dentistry does not work like that. A physician's day is a queue of appointments. A dentist's day is a chair, running two sessions, with a toothache walk-in that has to fit somewhere. A physician records a visit. A dentist records a stage: root canal on 36 today, obturation next week, crown after that, and eighteen months of braces adjustments for the patient in the next chair.

Dentisly is my attempt to build software around that day instead of around a generic visit. You can look around the live demo at dentisly.samikhanapps.com. This note covers why it exists, how it works, how I am building it, and what is honestly not finished.

The problem: a dental day has handoffs, not appointments

Watch one patient move through a clinic and you count four handoffs. The front desk books them. The assistant prepares the chair and takes vitals. The dentist treats and records. The counter bills them and books the follow-up. Every one of those handoffs is a place where something gets lost: the X-ray that lives in someone's WhatsApp, the discount nobody wrote down, the half-finished root canal that quietly drops off the book.

So the design question was not "how do I store patient records?" It was "how do I make those four handoffs lossless?" Everything in Dentisly follows from that.

Who it is for

Dental clinics, from a dentist running one chair alone to a multi-chair practice with associates, assistants and a busy front desk. It also suits orthodontic practices, which follow braces and aligner cases over many months and need to show a patient and their parents how the teeth have actually moved.

It is built in Pakistan, but it is not only for Pakistan. Plans are priced per region in rupees, in Saudi riyals and UAE dirhams, and in US dollars — set for each region rather than converted from one list price, because a number that makes sense in Karachi does not make sense in Dubai. A Gulf compliance pack for Saudi and UAE clinics, covering e‑invoicing and insurance claims, is planned, and so is Arabic, right to left.

It is worth being straight about the gap that leaves. The pricing reaches three regions; the software does not yet. Fees are held in rupees and the app runs on one clinic time zone, and neither is something an admin can change today. That is a real piece of unfinished work, and I would rather name it here than let a currency switcher on a pricing page imply the product is further along than it is.

How it works

Chair schedules that match reality. Each dentist gets a weekly chair schedule you set up once: Monday to Thursday, 10:00–14:00 and 17:00–21:00, Friday evenings only, Sunday off. A split morning and evening day is how a great many clinics actually run, so two sessions is the default rather than a workaround. There are emergency slots held back for toothache and broken-tooth walk-ins, so an emergency does not wreck the afternoon. Date overrides block Eid, public holidays and a dentist's leave without rebuilding the week. The front desk books into real gaps, which is what stops a slot being double-booked.

The whole chart, chairside. One profile per patient holds the clinical notes, the treatment plan and which stages are done, vitals recorded before anaesthetic or extraction, prescriptions, and the images. There is a tooth chart covering every tooth and surface with its treatment status, which is the part a general-medicine product never has. By the time the patient sits down, the dentist can see what happened last visit and what comes next. For ortho cases the images are kept in order visit by visit, so progress since bonding is something you can show rather than describe.

Your procedures, your fees. The clinic builds its own catalogue, from scaling and polishing through composite fillings, root canals, crowns, bridges, dentures and braces visits, each with its own fee in the clinic's own currency. Those fees then fill themselves in on appointments, treatment plans and invoices. Change a fee once and it applies from then on.

Billing that closes the visit. The invoice is built from the procedures actually done, with a per-line discount available on any item — a family discount on scaling, a goodwill discount on a re-do. The patient can pay part today, and the balance due is saved against both the invoice and their profile. That matters when the work is a root canal plus a crown, or eighteen months of braces, because the balance has to still add up on the fourth visit.

Five roles, because a clinic is not one person

Everyone gets their own login, and a role: Admin, Doctor, Nurse or Assistant, Front Desk, and Billing Officer. Each role only sees the screens and actions its job needs. A receptionist does not edit clinical notes. Only an admin manages staff accounts and clinic settings. Deleted records go to a recycle bin an admin can restore from, rather than vanishing. There is an audit trail of who changed what and when, accounts lock after repeated wrong passwords, and every request runs over HTTPS.

I want to be careful about how I describe that last part. Those are real protections, and they are not the same thing as a compliance certification. Dentisly does not claim to be HIPAA, GDPR or PDPL compliant, and I would rather say so plainly than let the word "secure" do work it has not earned. Login history, daily encrypted backups with tested restores, in-country storage, and patient consent and data-request tools are all on the list and not built yet.

How I am building it

This is the part other builders might find useful, because Dentisly is not a greenfield project and I decided early not to pretend otherwise.

I already had a working dental EHR — MEDICODENT, built and running. Dentisly is a rebrand and a hardening of that product, starting from the real exported source rather than a fresh rebuild. That choice came with two standing rules I wrote down for myself and have not broken.

The first is pixel parity. Every screen except the marketing home and login pages has to match the live app exactly — fonts, spacing, density, colours. When the local build differs, the live app wins and I fix the local build. The temptation with an inherited codebase is to redesign it while you are in there, and redesigning is how a two-week hardening job becomes a six-month rewrite that ships nothing.

The second rule is that the live app beats my own spec. When a task description disagrees with what the software actually does, I do not build to the spec. I log in, click through, read the exported component, build to match what is really there, and then correct the spec. Specs drift. Working software does not. That rule has saved me from "fixing" several things that were never broken, and it has forced a few honest upward revisions of scope that I would otherwise have argued my way out of.

The stack is React and TypeScript on the front end with a Node.js backend, Prisma and MySQL for data. I am both product owner and developer on it, which is a useful seat precisely because there is nobody to hand an awkward trade-off to.

Where it stands today

I would rather you heard this from me than found it yourself.

  • The Smart Assistant is not switched on. An AI helper for lookups and routine writing is designed and labelled, and it is off on the live site. So are voice-to-SOAP notes, a chairside chart summary, prescription safety checks, ortho progress drafting and no-show follow-ups. They are planned, which is a different word from available, and I am keeping them tagged that way until each one ships.
  • Invoices print, but they do not export. You can print an invoice for the patient. There is no PDF or Excel export of it yet. Doctor schedules do export to both.
  • There is no camera capture. Images are file uploads today, not a chairside shot from the device.
  • The Group plan is labelled planned, not available. Free and Clinic are live. Group, for practices running several clinics under one sign-in with a group dashboard, is on the pricing page with a price and a "planned" tag, because I would rather publish the intended price early than let a pricing table imply something ships today. There is also a student and new-graduate program for anyone opening a first clinic.
  • A good deal of the roadmap is still roadmap. Consent forms with e-signature, refunds and credit notes, booking by chair as well as by dentist, day/week/month views of the book, WhatsApp and SMS reminders, online payments and Arabic are all published as planned. Each one flips to available only when it ships.
  • The demo is a demo. It holds fictional patients only, caps uploads, and wipes itself every night. Please do not type real patient information into it.

When the AI does arrive, the rule it has to live by is already decided: it drafts, a person checks, and the dentist decides. Nothing clinical gets saved and no message reaches a patient without someone approving it. AI suggestions support clinical judgment; they do not replace it, and a product handling patient records is the wrong place to be relaxed about that distinction.

What I am treating as the hard part

The screens are the easy part. The hard part is the arithmetic and the rules underneath them, because that is what people rely on months later.

Three things in particular. Invoice maths has to survive a per-line discount, three part payments and a tax line and still produce a balance both the patient and the clinic agree on. Slot rules have to decide what "free" means when a weekly schedule, an emergency slot and a holiday override all apply to the same Tuesday. And role permissions have to be enforced in one place, not re-checked per screen, because the fifth screen is where someone forgets.

I keep those as shared modules rather than scattered logic, for a boring reason: a discount rule that exists in two places will eventually disagree with itself, and then the clinic stops trusting the balance. Once staff stop trusting the number on screen they go back to the register and the WhatsApp thread, and the software has failed no matter how good the charting is.

What I would like feedback on

If you run or work in a dental clinic, I would value your honest view, especially on the parts I may have simplified. The question I keep coming back to is this: which handoff in your day loses the most information? If you know the answer immediately, that is the thing I should build next.

You can try it at dentisly.samikhanapps.com, find me on LinkedIn, or see the rest of my work at samikhanapps.com.

More

Keep reading

All notes are on the blog index, and the shipped work is on the projects overview.