What's New in Serelora v1.7.0

This is the biggest release we've shipped, and almost all of it comes from one question we kept hearing in different forms: can the chart just tell me what's true right now?

Here's what changed.

The chart now knows the difference between "we're sure" and "it's current"

These are two different questions, and until now Serelora only tracked one of them.

Serelora is built on claims, not facts. Every piece of information in a chart is traceable back to where it came from, along with how confident we are in it. That's what makes a Serelora chart auditable in a way most systems aren't.

But confidence isn't the same thing as currency. We're confident this patient was diagnosed with asthma and this patient has asthma today are separate facts.

A condition or medication carries its own clinical status (active, resolved, inactive) alongside our confidence in the underlying record. That status arrives automatically from the records you import, and any provider can correct it directly from the chart in one click.

One deliberate choice worth naming: when the source material doesn't say whether something is still current, we say we don't know. We don't guess. A blank is more useful than a confident wrong answer.

A chart you can actually navigate

If you've ever opened a chart on a long-standing patient and been handed several hundred undifferentiated rows, this release is for you.

• Everything is grouped and sorted. Each chart tab now organizes rows into meaningful sections instead of one flat list, with clinically active items first.
• Filters that make sense. A single toggle shows just what's clinically relevant, or opens things up to include everything on record.
• Surgical History is its own tab. Procedures are classified so an appendectomy doesn't sit in the same undifferentiated list as a flu shot — and surgical entries now carry the surgeon, facility, indication, and any noted complication when the source records them.
• Lab panels stay together. A comprehensive metabolic panel arrives as a panel, not as fourteen unrelated results. Each analyte shows whether it moved since the last draw and by how much.
• Social History and Care Team have a home. Both were being captured and had nowhere to appear. They do now.
• Insurance reads like insurance. The coverage on file leads, followed by what the patient reported and what we found in documents — clearly labeled, in that order.

On lab trends, we show direction and change, but we don't label a result high or low. That would require a reference range we don't yet have, and inventing one is exactly the kind of quiet guess we don't want in a chart.

You can also add a condition or medication by hand now, for when you learn something in the room that isn't in any document yet.

Write your note by talking

The note editor now has a dictation microphone. Turn it on and your words appear in the note at the cursor, live, using medical-grade transcription. It works in a new draft or in an addendum to a note you've already signed.

Tables work properly now, too. When your assistant composes a lab panel or a medication list as a table, it lands in the editor as a real, editable table — not as a wall of pipe characters you have to clean up before signing.

An assistant that remembers what it did

Two changes make the in-chart assistant markedly more useful.

It remembers its own work. Previously, every turn started fresh: it could write a note or propose a change, and by your next message it had no memory of having done so. Ask it to "change that back" and it couldn't. Now it carries its recent actions with it, including whether you approved or declined them — so it never claims it made a change you actually cancelled.

It can propose record corrections and addenda. Tell it a patient's phone number changed, or that a signed note needs an addendum, and it prepares the change for your review. You see the current value and the proposed value side by side, and nothing is written until you approve it. The assistant goes through exactly the same audited paths your own edits do — it has no special access, and it can't sign anything.

Reminders that don't nag

Patients now get appointment reminders by both text and email: a confirmation when the visit is booked, a nudge when check-in opens a week out, and reminders two days and one day before.

The part we're most pleased with is the part that's quiet. If a patient has already completed their pre-visit check-in, the reminders adapt — they get "you're all set" instead of another request to do something they've already done. Changes and cancellations send their own notice, and un-cancelling a visit tells the patient it's back on rather than pretending it was booked fresh.

Every patient can control this. A new Message settings page in the patient portal lets them choose which channels they hear from, and your staff can set the same preferences from the chart. Appointment reminders and registration invitations are controlled separately, so a patient who silences visit reminders still receives the intake link your front desk needs delivered.

Clinics get a switch of their own, so you can turn appointment notifications on for your practice when you're ready rather than all at once.

The schedule reads at a glance

Visit types now carry both a color and a shape on the calendar, so you can see what your day looks like without opening anything. The shape matters: color alone isn't accessible to everyone, and a schedule is something you scan under pressure.

Your practice defines its own visit types — the labels, the order, and the styling — from an admin page.

We also fixed something that could quietly hide work from you: on a very busy week, appointments past a certain count simply weren't drawn, and the affected days looked empty. They no longer do.

Money, at the moment you can collect it

The front desk is where cash-pay balances actually get paid, and check-in gave staff nothing to work with. Now the check-in dialog shows the patient's outstanding balance and lets your team take payment right there.

Patients see the same number. The portal's billing tab now leads with what they owe across all their visits, so they aren't left adding up invoices themselves.

A patient list that doesn't slow down

For practices with large panels, opening the patient list meant waiting on the entire roster — every patient, every time. It now loads in pages as you scroll, and search runs on the server, so the list stays quick whether you have two hundred patients or ten thousand. The appointment patient picker got the same treatment.

Browser Back behaves properly now, too: return to the list and you land where you left off, with your search intact, instead of at the top.

Getting your data in

If you're coming from Office Ally, you can now import your patient roster directly from a Practice Mate export. Serelora previews every row before anything is written, tells you which patients are new and which look like possible duplicates, and never auto-merges a match — that decision stays with you. Large rosters import in the background so you're not stuck waiting on a browser tab.

Records arriving from other systems are also being read more carefully. Demographic details — marital status, race and ethnicity, employer, language and interpreter needs, emergency contacts — now fill in from incoming records when the corresponding chart fields are empty. Fields you've already filled in are never overwritten.

We also rebuilt how we read C-CDA documents. A single "Asthma, active, confirmed" entry used to become three separate condition rows; now it becomes one, with the status attached where it belongs. In our test documents, the same file produces about a third as many rows — the same clinical content, without the clutter.

Privacy and safety

A few changes we think are worth stating plainly.

No third-party scripts run on a patient chart. Any script loaded into a page can read what's rendered on it. We've removed the last one from the provider application and added an automated check that fails our build if one is ever reintroduced. This is now a standing rule, not a preference.

Chart numbers identify one patient. Medical record numbers are now compared without regard to letter case, so ABC-1 and abc-1 can't become two charts for the same person within a practice.

Error messages don't leak. We audited every place the system reports a problem and made sure diagnostic detail stays in our logs rather than reaching a screen or an error message — including some deep internal paths where database errors could echo back the record being written.

Documents are matched to the right patient. Documents that don't name the patient they were filed under no longer contribute clinical information to that chart. The document is still stored and readable — it just doesn't quietly become part of someone's medical history.