vs Atlas — /vs-atlas/
A single-page, marketing-grade head-to-head between Starlight.MD (2026) and Atlas.MD (2014). Hero stats, 30-second pitch, 10 numbered dimensions, a side-by-side chart screen, and bottom-line ROI for both Yogini specifically and the broader DPC + concierge category.
Atlas is an EMR built in 2014. Yogini's parents text a phone number. Yogini charts after dinner. We built the EMR for 2026: AI scribe substrate, native parent app, family billing, newborn home visits as first-class. Below: 10 dimensions, side-by-side, with numbers.
Hero stats
| Stat | Value | Meaning |
|---|---|---|
| Evening charting recovered | 30–60 min/day | ~$30K/yr of doctor time freed |
| Sign → parent push | <60s | From "Sign" tap to push notification on Lisa's phone |
| Dimensions Atlas loses on | 10 | Each with numeric proof |
| Insurance / claims / RCM build | $0 in v1 | We never build it. DPC doesn't need it. |
The 30-second pitch
Atlas.MD · 2014. A desktop EMR. Phone number. Manual everything. Built when iPhones were 4 years old and Stripe was 3. Server-rendered jQuery. Parents text a Twilio number and wait. The doctor types notes after dinner.
"this feature is not available on mobile" — Atlas tutorial tooltip captured in Yogini's account, 2026-05-04.
Starlight.MD · 2026. An AI-substrate EMR with a parent app. Native everything. Built when Claude can transcribe + summarize a visit in 20 seconds. The chart still has the single timeline (Atlas got that right). Everything around it is rebuilt for how care actually works in 2026.
The 10 dimensions
| # | Dimension | Atlas | Starlight | Bottom line |
|---|---|---|---|---|
| 1 | Visit notes | Type by hand, often after dinner. Dragon dictation on top doesn't structure into SOAP. | Tap "Start visit · AI scribe". Whisper + Claude draft SOAP from transcript + chart context. Sign in ~30s. | +$30K/yr in Yogini's evenings recovered, or 2–3 more visits/day. |
| 2 | Parent communication | Twilio number → unified inbox. No native app. No push. No per-kid threading. | Native iOS + Android via Expo. Push, per-kid threads, in-context refill, photo upload, payment, family switcher. | The mobile-app moat. Once installed, parents don't switch back. |
| 3 | Visit summary for parents | Nothing automatic. Yogini types it back via SMS if asked. | 2nd Claude pass auto-generates plain-language summary. Push within 60s of sign. | The leapfrog. Less "what did the doctor say?" follow-up. Higher NPS. |
| 4 | Newborn home visit | Generic SOAP template. Birth history as ad-hoc free text. | Dedicated Home Visits tab. Birth history pulled from intake + hospital discharge. Tap-chip measurements on iPad. Auto-1-week well-baby on sign. | Premium-priced differentiator. "Same-day home visit for your newborn" — Atlas can't deliver it. |
| 5 | AI | One sidecar feature ("Gap Analysis"). Doesn't write notes, doesn't triage, doesn't draft replies. | AI is the substrate. Scribe, message triage, risk panel, drug-interaction lookup, parent summary, doc auto-routing, prior-auth letter draft, refill auto-drafts. Anthropic BAA. | Atlas can't catch up without rebuilding. Each Claude release makes Starlight better. |
| 6 | eRx | "Go to Rcopia" SSO redirect. Search drug, fill context manually. Two products taped together. | Native eRx modal in chart action bar. Quick favorites pre-loaded. Patient context auto-filled. DoseSpot under the hood. | 30s vs 2 min per Rx. 15–20 min/day saved at 8–12 Rx. |
| 7 | Documents | Faxes inbox + files inbox (separate). Manual triage. Forms are blank PDFs. | Unified Documents tab. Inbound auto-routed by Claude (name+DOB extraction, 79% hit rate). Outbound forms pre-filled from chart. | Staff stops being a filing clerk. |
| 8 | Family / household billing | Each kid is a Stripe customer. Lisa's two kids = two billing relationships. Shared custody is a workaround. | Household first-class. One Stripe customer per family. Family-plan pricing ($295/mo for 2 kids vs $165/mo for 1). Split-billing flag. Employer billing for corporate-sponsored DPC. | This alone wins us pediatric DPC. |
| 9 | Mobile | Desktop-only. Tutorial flag literally says "this feature is not available on mobile." Yogini takes home-visit notes on paper. | iPad-first responsive. Touch-optimized chips for newborn measurements. Native iOS + Android parent app. | Home visits + after-hours messaging actually work in 2026. |
| 10 | Reports + practice management | 7 hard-coded billing widgets. No clinical-quality, no custom queries, no HEDIS-lite. | Same 7 in v1 (Atlas parity). Plus clinical-quality (HEDIS-lite, well-visit completion, immunization compliance, no-show rate, panel management). Custom query builder in v1.5. | Practice management based on data, not gut. |
Side-by-side chart screens
The same patient (Sofia Martinez · 7y · peanut allergy · cough × 4 days) rendered in both products. Atlas's chart is a chronological flat feed of one-line entries; Starlight's chart is a hero card + AI risk panel ("Well-visit overdue · Epi-pen expires 47d · Mom texted 4d ago about cough → auto-triaged") with clickable ICD codes, parent name, sibling, and pharmacy.
Bottom line — for Yogini's practice (81 patients today, ~$57K ARR)
| Lever | Lift | Why |
|---|---|---|
| Doctor time recovered | +$30K/yr | 30–60 min/day on charting via AI scribe |
| Family retention | +10–20% | Mobile-app families churn less than SMS-only |
| New family conversion | +15–25% | Self-serve magic-link onboarding + first-visit-booking-in-flow |
| Newborn home visit | +1 segment | Premium add-on (~$200/visit) or $50/mo plan tier |
Payback on doctor time alone: 2 months.
Bottom line — for the DPC + concierge category
| Metric | Value |
|---|---|
| DPC EMR TAM (today) | ~$50M (5,000 US DPC practices × ~$10K/yr) |
| Adjacent TAM unlocked | ~$200M (concierge + larger pediatric DPCs) |
| Practices in 12 months | 10–15 (Yogini is pilot 1) |
| Atlas catch-up time | ~3 yrs (rebuild stack + ship parent app + rewrite AI surface) |
How Deepak knows what Atlas does
Every claim is sourced from a logged-in tour of Yogini's actual starlightpediatrics.atlas.md account: 33 screens captured, 15 chart-internal tabs mapped, 8 third-party integrations identified.
The Atlas screens contain real PHI. They're scraped and live only on Deepak's laptop in
~/Desktop/starlight-md/atlas-scrape/— gitignored, never pushed.
Identified Atlas integrations: DrFirst Rcopia (eRx) · Rubicon MD (eConsults) · Stripe · Twilio-likely · NCPDP/Surescripts · NDC · ICD-10 · SMTP/OAuth email.
Practice context for Yogini's Atlas account: 81 patients · +20.9% YoY · $4,752 May charges · 100% Stripe · 1 doctor · 0 in-house dispensing.