- README: step-by-step Google Cloud Vision API key creation and setup
(enable API, create service-account key, place as google_credentials.json)
- ocrLocate: split into getOcrWords()/matchWordsForText() so a screenshot is
OCR'd once and matched locally against multiple text targets, instead of a
separate billed Vision API call per target (patient-row and Exam steps each
cut from 2-3 calls down to 1)
Claude vision's self-reported text coordinates had a small but consistent bias,
causing the patient-row double-click to land above the name and the Exam-column
divider count to occasionally miscount and hard-fail. Swap patient-row, Exam,
and Save button targeting to exact OCR bounding boxes (Google Vision via the
existing PaymentOCRService) instead of AI-estimated ratios.
Deleting/replacing an appointment attachment left an orphaned ClaimFile
copy behind (auto-linked earlier by autoLinkAppointmentFiles), causing
the Selenium claim automation to upload extra files beyond what's
currently attached to the appointment. Claim PUT was also append-only
for claimFiles, so removing an attachment in the UI never persisted.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Row selection previously computed the target row's y arithmetically
(header.y + rowHeight * rowIndex), which drifted further off-row the
deeper the target patient was in the filtered list, and an AI
"onTarget" verification step added to catch drift instead confirmed a
click point that was visibly in blank space below the grid.
locateAllOnScreenshot now reports each text match's top/bottom edges
instead of a self-estimated center, and the patient row step locates
the target's first and last name directly, pairing whichever
occurrences land on the same row instead of trusting arithmetic or a
qualitative AI judgment call.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Replaces fragile whole-screen text matching with a screenshot-diff step that
finds each dialog's actual pixel bounds, then restricts every subsequent AI
locate/click to that cropped region — eliminating false matches from text
elsewhere on screen (title bars, side panels). Adds column-boundary and
row-height detection so the patient row and Exam procedure click positions
are computed geometrically instead of relying on repeated fuzzy AI guesses
for visually similar neighbors. Also adds per-run screenshot/debug-crop
backups and a structured run.log for diagnosing failed runs, plus a
cmd:move primitive on the Windows agent for pre-click confirmation crops.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds a Windows-side agent (screenshot/click/type over HTTP) plus backend
services to locate UI elements via vision and drive an existing-patient
appointment flow in Open Dental, wired into the Copy/Type Agent page and
socket progress updates.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- ccaPreAuthProcessor: keep claim status as PREAUTH after successful
submission instead of jumping straight to APPROVED, matching the
Tufts SCO/UnitedDH/MassHealth preauth flows; also tag PNG fallback
screenshots with image/png instead of the default application/pdf
so they render in the Documents preview modal.
- claims-recent-table: add a PREAUTH status case so it displays as
"Submitted" instead of falling through to the raw enum value.
- Chatbot: submitting with only a screenshot attached (no text) now triggers the same
"AI Detect & Check Eligibility" flow as the Copy Agent page
- detect-eligibility-info now also extracts the visible insurance payer name and picks
Tufts SCO auto-check when both Sun Life and DentaQuest are detected, else MassHealth
- Chatbot-attached files are now handed off to both claim and preauth forms unconditionally,
not just claims
- TuftsSCO preauth Selenium worker: verify-and-retry the Tooth field (typing was silently
getting reset by duplicate-procedure-code warning banners), add a final re-verification
pass across all rows, fix acknowledgement-checkbox targeting/verification, and verify
the "Next step" click actually advances the wizard instead of trusting a blind click
- selenium-claim route no longer re-appends previously saved appointment
attachments when the chatbot's autoSubmit flow already sent every intended
file, preventing an unwanted extra upload (e.g. attach 2 -> Selenium sees 3).
- Add a "Take screenshot" button to the AI chatbot's attachment bar, reusing
the same capture/crop flow as the appointment editor, so claims and
preauths can be submitted from the chatbot with a live screenshot attached.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- Add "surg ext" alias mapping to D7210 in the AI CDT lookup.
- Add GET/DELETE endpoints for appointment attachments, and load existing
attachments in the appointment edit modal (they were uploaded correctly
but never re-fetched, so they appeared to vanish on reopen).
- Fix claim-form.tsx so previously-saved appointment attachments are always
loaded (even when a claim already exists) instead of only when creating a
brand-new claim, preventing them from being silently wiped on save.
- Merge existing appointment attachments into every insurance submit handler
(MassHealth, CCA, DDMA, UnitedDH, TuftsSCO) so Selenium submission always
includes files saved earlier via the Schedule editor, not just files
uploaded in the current claim-form session.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds an optional Provider (NPI) select to the Edit Appointment form,
sourced from Settings > NPI Providers. The AI "Claim for Column" batch
flow now prefers the appointment's own provider when resolving which
provider Selenium should select on the insurance site (e.g. MassHealth
rendering provider dropdown), falling back to the existing per-procedure
selection, active claim, or first configured provider.
Also resyncs packages/db's committed compiled schema/client artifacts
with their .ts sources — they had drifted out of sync (some untouched
since July), which was silently stripping the new npiProviderId field
via stale .strict() zod schemas.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
MassHealth eligibility lookups via the AI chatbot only matched existing
patients on memberId+DOB, so patients imported from PDF extraction (which
have a member ID but no stored birthday) were treated as new/unknown even
though they already exist in the system.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Lets staff attach files to an appointment directly from the edit form via
local disk upload, picking a previously-saved scanned attachment, or an
inline screenshot with crop-before-save (matching the Copy Agent page's
crop UX), instead of detouring through the Copy Agent page. Also fixes a
latent data-loss bug in claim-form's save-for-appointment call, which
replaced all AppointmentFile rows instead of merging with existing ones.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- AI claim queue now uses previously saved AppointmentProcedure data (with a derived insurance site key) when an appointment has no notes but has saved procedures, instead of skipping with "No notes on this appointment"
- selenium-claim route now reads saved AppointmentFile attachments from disk when they weren't re-uploaded in the browser, so images/PDFs already attached via "Select Procedures" reach the MassHealth submission
Adds an "AI Detect & Check Eligibility" action to the Copy Agent page that
sends captured screenshots to Claude to extract Member ID/DOB, then opens
the chatbot's eligibility confirm step pre-filled with the result.
- POST /api/ai/detect-eligibility-info: Claude vision extraction of
Member ID/DOB from uploaded screenshots.
- POST /api/patients/create-temp: auto-creates a placeholder "PID-<id>"
patient (matched by insuranceId+dob) when no patient is selected, so
screenshots can be saved immediately; the real eligibility check later
rewrites the patient's name and Cloud Storage folder automatically via
the existing createOrUpdatePatientByInsuranceId matching logic.
- Chatbot: Member ID/DOB on the confirm card are now editable (MM/DD/YYYY),
with edits synced back to the patient record before the check runs.
- Fixed a useEffect ordering bug on the insurance-status page that could
silently clear a chatbot-prefilled Member ID/DOB before the auto-trigger
effect ran, preventing the eligibility check from ever starting.
- Added a local-disk-only screenshot backup (apps/Backend/uploads/
screenshot-backups/) for debugging, not exposed anywhere in the UI.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Native file inputs can't be pointed at a default server folder for security
reasons, so instead the Browse buttons on the Claim/PreAuth/Select Procedures
forms and the AI chatbot's attach button now also offer picking from
previously saved Cloud Storage attachments (e.g. Copy Agent screenshots) via
a new server-side picker modal, scoped per-patient on the claim forms and
unscoped (recent across all patients) in the chatbot.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
getDisplayMedia (screen capture) requires a secure context, so LAN
staff access now goes through nginx on a real trusted cert
(local-summit.mydentalofficemanagement.com, issued via certbot's
Cloudflare DNS-01 plugin) instead of plain HTTP. A separate public
hostname/nginx block is reserved for Twilio webhooks only, routed
through Cloudflare Tunnel so nothing else is exposed to the internet.
Also fixes the backend CORS allowlist for the new hostname, and stops
the Socket.IO client from bypassing nginx to hit the backend directly
on :5000 (which broke under TLS/mixed content).
filesForQueue was collected from AppointmentFile/ClaimFile records but only
wired into the MH and CCA job payloads. The other three insurers' Selenium
workers read claim.claimFiles and always got an empty list, so attachments
already saved to Cloud Storage silently never got attached to the claim.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- PDF viewer hung on "Loading" forever for cloud-backed PDFs: streamFileTo
pipes with {end: false} (by design, so cloud-storage.ts's own routes can
call res.end() themselves), but the new /pdf-files/:id cloud branch never
did, so the HTTP response never completed.
- Documents page showed nothing for patients whose PDFs are only in Cloud
Storage: the page lists categories from the legacy PdfGroup table, which
new saves stopped populating. savePdfToCloudStorage now also ensures a
blob-less PdfGroup row exists per patient/category so the category still
appears, with actual files served from Cloud Storage via the existing
merge in the pdf-files/group listing routes.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New saves go into the patient's Cloud Storage folder under category subfolders
(Eligibility/Claims/PreAuth/Claim Status/Attachments) via a new
storage.savePdfToCloudStorage helper, instead of the legacy PdfGroup/PdfFile
Postgres blob tables. Old records and their read paths are left untouched for
backward compatibility; PDF list/content/delete endpoints and viewers now
transparently handle both legacy numeric ids and new "cloud:<id>" ids.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Renames the sidebar entry to "AI Copy/Type Agent" with Copy Agent, Type
Agent, and Settings sub-pages. Settings shows the server's LAN IP
address (via a new /api/ai/network-info endpoint) since it changes
depending on which office the app is running in.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claim attachments were being appended without checking for existing
duplicates, causing files (e.g. the same PDF) to be sent to the
insurance clearinghouse multiple times.
Adds a new Settings → Advanced → Selenium Settings page where the office
name (paymentGroupId) used in United/DentalHub portal dropdowns can be
configured without touching code. Previously hardcoded as "Summit Dental
Care"; now injected from DB into the three United workers at job dispatch.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New lab_rx_template table with sortOrder and Prisma migrations
- Backend CRUD + reorder routes at /api/lab-rx/templates
- Auto-seeds 3 Highland Dental Studio defaults on first use per user
- LabRxModal: template picker, inline rename, up/down reorder, full
template form with sections (Template Name / Lab Info / Case Details)
- RX popup: Tooth Number, Shade, Due Date, Doctor (from NPI providers),
Instructions pre-filled with "Please make / Thank you!", print output
- lab-rx-templates.ts config file for future git-managed additions
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The classifier prompt only told the AI that appointmentDate applies to
schedule_appointment/claim_only/check_and_claim, so a trailing date like
"all on 6/23/26" was dropped for multi-patient claims even though the
workflow already reads c.appointmentDate for those intents and silently
defaulted to today.
Also add Prisma config/seed scripts and shopping-vendor types/schema updates.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add multi_claim intent so AI correctly handles "claim X for patient A and Y for patient B"
instead of applying all procedures to all patients (batch_claim)
- Each patient carries their own matchedCodes in the queue
- Fix batch claim PDF race condition: chatbot queue no longer advances in closeClaim(),
instead advances after selenium PDF is downloaded (matching column claim behavior)
- Align United SCO eligibility worker with claim worker: only fill subscriber ID + DOB,
use treatmentLocation by ID instead of arrow-wrapper click
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add "recement" example to LLM classifier so it spreads across multiple
teeth (recement #5, #3 → two D2920 entries). Add "today" to the AI
column claim message so the LLM sets appointmentDate instead of asking
"Which service date?"
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Make appointmentId nullable/optional in Prisma Zod schema via @zod rich
comment so claims can be submitted without an existing appointment
- Convert undefined appointmentId to null in all claim form handlers and
the backend claim creation endpoint
- Add AI classifier rule for expanding one procedure across multiple
comma-separated tooth numbers (e.g. "post/core on #23, 24, 25, 26")
- Add "post/core" (slash) alias to CDT lookup maps
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add "comp" as alias for composite in CDT lookup. Update classifier
prompt with explicit examples for "claim comp #8 ml for lisa today"
and dental surface letter definitions (M/D/L/F/B/O/V/I) so the LLM
correctly treats #tooth+surfaces as composite fillings, not insurance.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a toggle behind the Go button on the Payments page to automatically run the MH batch payment check on a user-selected day of week and hour. Default is off.
- Schema: added autoMhCheckEnabled, autoMhCheckDayOfWeek, autoMhCheckHour to User model
- Backend: new mhBatchPaymentService (shared logic), GET/PUT /auto-mh-check-setting routes, hourly cron job that fires on matching day+hour
- Frontend: toggle + day select (Mon–Sun) + time select (hourly) that persist immediately to DB
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The auto-import wipes the database, destroying the cronJobLog record
created at the start of the job. Now duration is calculated from a
local variable instead of querying the DB, and if the record is gone
a fresh log entry is created in the restored database.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- PDF import now marks payments as PAID when MassHealth patient's
mhPaidAmount >= totalBilled (no patient balance)
- Newly created patients from MH vouchers get insuranceProvider = 'MassHealth'
- Existing patients with blank insuranceProvider get it filled on import
- Fix: update patient name from PDF if existing record has empty name
- Various frontend/selenium/route updates
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
applyMissingMigrations ran the entire migration SQL as one transaction,
which rolled back all changes when any statement failed (e.g. adding a
NOT NULL column to a non-empty table). Replaced with prisma db push
which compares the schema and applies only what's needed. Also added
support for importing .sql files (not just .zip) in auto-import.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
WebDAV server now serves the Backend root directory. Receiver pulls
backups/, chat-history/, and uploads/ individually so all data is
synced between PCs.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds autoImportEnabled/autoImportHour to rclone config. A separate
hourly cron finds the latest .zip in backups/ and restores it to the
database (drop schema, psql restore, apply migrations). Frontend shows
toggle + time picker + Import Now button in the Receiver PC section.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace hardcoded 8 PM / 9 PM backup schedules with user-selectable
hour dropdowns. Adds autoBackupHour and usbBackupHour fields to User
model. Cron jobs now check every hour against the configured time.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Source PC serves backups/ folder via rclone WebDAV server (auto-starts with app).
Receiver PC pulls backups on schedule using rclone sync.
Network Backup UI now has two tabs: Rclone and API Key.
Rclone installed automatically via postinstall script.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Requests from outside 10.x, 192.168.x, 172.16-31.x, 127.x are rejected
with 403 before the API key is even checked. This prevents the database
dump endpoint from being reachable from the internet even if the key leaked.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The /network-backup and /network-backup-files routes use their own
X-Network-Backup-Key authentication. Mounting them behind authenticateJWT
blocked all receiver sync requests before they could be validated.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Copy-paste often adds trailing whitespace that breaks the API key comparison.
Updated placeholder and description to show the correct URL format (no port number).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add parseMultiPaCode: "4PA #3#14#19#30" expands to D0220 + 3× D0230 with tooth numbers
- Add "3rd pa"–"6th pa" direct aliases → D0230 as fallback
- Tighten LLM prompt to always use "2nd pa" for all additional PAs, never "3rd pa" etc.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- appointments-page: save full queue from current patient on confirm so that
cancelling the claim form resumes at the same patient instead of skipping it
- claims-page: advance ai_claim_queue only after successful submission (CCA,
MH selenium PDF download, or direct non-draft submit) via new advanceAiClaimQueue()
- cdt-lookup: fix extractToothSurface regex to allow space between # and digit
(e.g. "# 32") so tooth number is correctly captured
- cdt-lookup: silently skip bare tooth-number tokens ("# 32", "#32") that the
LLM may emit as standalone items, preventing false "unrecognized procedure" errors
- internal-chat-graph: add explicit rule that a CDT code followed by # NN
(e.g. "D7210 # 32") must stay as one entry and not spread to other procedures
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add batch_eligibility_by_name intent so "check Mary and Sinthia" resolves multiple names
- Add "Check All & Appointment Today" button to batch eligibility UI — creates appointment for each patient after their eligibility check completes
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
