Company profile · 4 min read
CareSwift: agency-controlled EMS narratives
CareSwift drafts editable ambulance narratives against an agency’s documentation standards while providers retain review and finalization.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
What it does
CareSwift helps ambulance crews produce incident narratives faster and with an agency’s own documentation standards. The current workflow asks providers for call details, surfaces missing or unclear information, drafts an editable narrative, and leaves the provider responsible for reviewing and finalizing it. The YC profile describes a real-time assistant for ambulance reporting; the current homepage is explicit that the product works alongside an existing ePCR and that CareSwift does not write back or submit charts.
The fit is an EMS agency whose crews lose time to documentation, corrections and supervisor follow-up after calls. This is not a replacement ePCR. It is a focused narrative step that can be configured around agency-approved call types, questions and writing rules.
Why I’d look closer
The product’s boundary is its strength. Providers review, edit and finalize the narrative, then copy it into the agency’s ePCR. The homepage says the system does not certify that a narrative is clinically complete, billable or approved through QA, and that the provider remains responsible. It also says the company can measure adoption, time to finalization, abandonment and copy activity, while agency-supplied data can support comparisons of late charts, corrections and sendbacks.
CareSwift reports 100,000+ EMS narratives generated by providers and says the product is HIPAA compliant in an approved agency deployment. Those are company-reported claims. The homepage describes a required security review, BAA, provisioning and retention preset before PHI use; it also says there is currently no write-back, sync, supervisor approval, comments, sendbacks or correction queue in CareSwift itself. Those constraints should be part of the buyer fit, not hidden in a footnote.
The founder context is unusually direct. The YC biographies describe Brian Weigand as a current EMT and former NYC 911 ambulance worker who built ambulance-reporting software, and Jonathan Zero as the technical co-founder building AI ePCR tools.
What I’d ask
Which ePCRs, browsers and devices are supported, and how much manual copy/paste remains on a real shift? I’d run synthetic and de-identified narratives across several call types, inspect the clarification logic, and measure late-chart and correction rates against a baseline. Pricing is agency-scoped and demo-led, so implementation and security review should be in the same evaluation.
My editorial take
CareSwift is a strong fit for EMS agencies that want to reduce narrative burden without delegating clinical responsibility. The deliberate “does not write back” boundary is a real limitation, but also makes the first pilot easier to govern. The proof is a measurable reduction in late or corrected charts, not a generic AI-scribe claim.
Quick facts
| Field | Sourced detail |
|---|---|
| Buyer fit | EMS agencies and ambulance crews |
| Workflow | Clarify, draft, provider review/finalization, copy into existing ePCR |
| Public claim | 100,000+ narratives generated; HIPAA-compliant agency deployment; company-stated |
| Pricing | Paid agency product with scope-based pricing; no public price card |
Sources checked
Checked 2026-09-19.
| Source | Used for |
|---|---|
| YC company profile | Product, founders and launch workflow |
| CareSwift homepage | Current workflow boundary, security/PHI claims, limits and pricing posture |
Cohort context
CareSwift is listed in Summer 2025. In our 2026-09-18 directory snapshot, 11 of 166 listed companies in that cohort have YC’s primary industry label Healthcare (6.6%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:17:58.983Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗