AI for healthcare,built by physicians.

We advise, automate and build software for clinics and hospitals in Mexico, so your team spends less time on paperwork and more time with patients.

Book a callSee solutions
Fig. 01 · Astrocyte

  1. 1. We analyze

    We learn how you operate and find where AI actually helps.

  2. 2. We build

    Custom software, reviewed with your clinical team.

  3. 3. We support

    Rollout, training and continuous improvement.

The problem

The clinical information exists, but it can't be used: it's scattered across Word, Drive, paper and systems that don't talk to each other.

Document intelligence

The information is already in your documents. Our AI reads it, structures it and the physician confirms it.

progress_note.pdf
Progress noteOutpatient clinic

Patient: M. R. F. · 58 years · Record 004512

58-year-old female patient with a history of type 2 diabetes mellitus treated with metformin 850 mg every 12 hours. Follow-up visit. BP 150/95 mmHg, capillary glucose 182 mg/dL. Diagnosis: systemic arterial hypertension. Starting losartan 50 mg every 24 hours.

Dr. A. L. · Internal medicine

AI proposalReading scanned document

  • History

    Type 2 diabetes mellitus without complications

    E11.9

  • Diagnosis

    Essential (primary) hypertension

    I10

  • Medication

    Metformin 850 mg PO every 12 h

  • Medication

    Losartan 50 mg PO every 24 h

Pending confirmationConfirmed by the physician

Nothing enters the record without the physician's confirmation.

Sample data

One system

From Word, Drive, paper or a legacy system to one integrated record with structured data.

  1. Any format

    Word, PDF, Drive or paper documents

    18,420

  2. AI unification

    patients, with no duplicate records

    2,310

  3. Structured data

    diagnoses, medications and lab results

    41,876

  4. One health record

    for practices, clinics and hospitals

    live

Most frequent diagnoses

  • J06.9Acute upper respiratory infection412
  • I10Essential hypertension356
  • E11.9Type 2 diabetes mellitus298
  • E66.9Obesity187

342

First-time visits this month

128

Pregnancies under follow-up

Sample data

Solutions

What we build. For the consulting room, the hospital and its data.

01 / 04

Clinical AI

Models that listen to, read and structure clinical information. The physician always confirms.

  • Document intelligence

    Diagnoses, medications and lab results extracted from PDFs, Word or photos and coded in ICD-10, ICD-9-CM and LOINC.

  • Medical dictation

    The spoken consultation becomes a structured clinical note.

  • Evidence-based clinical assistant

    Ask during the visit and get answers cited from the record, PubMed, DailyMed and the FDA.

  • Clinical agents

    Ask in plain language, “find this week's pregnant patients” or “book an appointment”; the agent orchestrates its tools.

Clinical assistantM. R. F. · 58 years

Summarize this patient's record

58-year-old woman with type 2 diabetes mellitus treated with metformin.1 On 10/08, systemic arterial hypertension was recorded and losartan 50 mg every 24 h was started.1 Her latest blood count shows hemoglobin of 10.9 g/dL, below the reference range.2

  • 1Progress note · 10/08/2025
  • 2Complete blood count · PDF
  • Searches: record · PubMed · DailyMed · FDA
Sample data
02 / 04

Digital hospital

Your whole team on one record, with data flowing between departments.

  • Collaborative health record

    Every department records its patients in the same health record, with permissions.

  • Migration from any format

    Word, Drive, paper or legacy records, with no vendor lock-in.

  • Data flow between departments

    Information moves from one service to the next without re-entry.

  • HL7 FHIR integrations

    Your systems exchange records, reports and prescriptions.

Single health record4 departments
MRM. R. · 58 years

  1. Admission note08:14

    Emergency · Dr. A. L.

  2. Blood chemistry validated09:02

    Laboratory · C. S.

  3. Chest X-ray report10:37

    Radiology · Dr. P. M.

  4. Nursing orders11:20

    Inpatient · Nurse J. R.

Related evidence: 3 publications on the diagnoses recorded by the physician
Sample data
03 / 04

Quality and data

Consistent records, regulatory compliance and visibility across the whole operation.

  • Standardized notes

    The same clinical structure across the hospital.

  • Quality control

    Records checked against the consultation transcript.

  • Traceability and observability

    Who did what, when, and who viewed each record.

  • Business intelligence and research

    Continuous metrics and reliable lab data for research.

Traceabilitylive
Who viewed M. R.'s record today?
  • 08:14

    Created the admission note

    Dr. A. L. · Emergency

  • 09:40

    Read the admission note

    Dr. P. M. · Radiology

  • 11:20

    Recorded vital signs

    Nurse J. R. · Inpatient

  • 12:05

    Signed the progress note

    Dr. A. L. · Emergency

You can also ask

  • How many first-time visits did I see this month?
  • What are the most frequent diagnoses?
  • How many pregnancies am I following?
Sample data
04 / 04

Patients and operations

Automations that organize daily work and support patients.

  • WhatsApp automations

    Reminders, confirmations and document delivery.

  • Smart scheduling

    Self-booking, appointment types, connected calendars and video visits.

  • Pre-visit intake

    Patients answer before the visit, by voice or text.

  • Safe prescribing

    Drug interaction alerts when prescribing.

Today's schedule4
  • 09:00MR

    New patient

    Confirmed via WhatsApp

  • 10:30JL

    Follow-up

    Reminder sent

  • 12:00AT

    Video visit

    Zoom link sent

  • 16:30CP

    Self-booked

    Intake pending

Sample data

Services

Four ways to work with us.

01

Healthcare technology consulting

We help you decide which technology you need, and which you don't, before you invest.

  • Process and systems analysis
  • Electronic health record selection
  • NOM-004, NOM-024 and data protection
  • AI strategy where it actually helps
02

Healthcare automation

We take the repetitive, manual tasks off your team's plate.

  • Self-booking and pre-visit intake
  • WhatsApp reminders and documents
  • Clinical notes from voice
  • Reports for management
03

Medical software development

We build the system your institution needs when an off-the-shelf one doesn't exist.

  • AI agents with physician validation
  • Data extraction from lab reports
  • Health record migration
  • HL7 FHIR integrations, web and mobile
04

Ready-to-deploy products

Platforms already built and validated with physicians: deploy them without paying for development.

  • No development costs
  • Validated with physicians
  • Astrocyte: clinical AI workstation
  • Rollout and training
Explore our products

Our own APIs

Components ready to deploy. An ecosystem of validated APIs and workflows to automate your institution's processes.

Astrocyte API

Electronic health record

API

Connect the record to apps, reports and clinical studies.

Bring it to your institution

Coding and interoperability

Astrocyte Search

API

Semantic search and autocoding in ICD-10, ICD-9-CM and LOINC.

Bring it to your institution

BlinkPDF

PDF documents

APIMCP

Fill official forms and generate accessible PDFs.

Visit site

PuffinEarly access

Clinical and scientific documents with LaTeX

APIMCP

Publishing-quality documents with verified references.

Bring it to your institution

Neuroglia Inference

Specialized medical inference

API

Validated clinical calculators and specialized models.

Bring it to your institution

Record and document digitization

Astrocyte · Tailored

Service

From paper, PDFs or legacy systems to structured data.

Bring it to your institution

Interoperability

Connected where your patients are.

WhatsApp, calendars, video visits and clinical standards: information reaches wherever your team and patients need it.

Integrations4

  • WhatsApp

    Reminders, intake and documents

    Active

  • Google Calendar

    Availability and schedule blocks

    Active

  • Zoom

    Video visits from the schedule

    Active

  • HL7 FHIR

    Exchange with other systems

    Active
ZDR
zero data retention AI models
ICD-10
ICD-9-CM and LOINC with semantic search
FHIR
HL7 standard for interoperability
NOM-024
and NOM-004 built in

Cases

Custom development for healthcare institutions. From many tools and formats to one system.

End-to-end automation01

An online clinic, from many tools to a single platform

Online consultation clinic

Online forms, Calendly, Google Drive, handwritten notes and PDFs and personal WhatsApp became one automated end-to-end workflow, with traceability and business intelligence.

Integration02

An enterprise API for their patients' app

Digital health company

We connected their app to the health record through an enterprise API: what physicians record during the visit goes straight to the app, patients see their information in real time, and it all plugs into their logistics.

Migration03

From any format to structured data

Practices and clinics

Unstructured notes, PDFs, lab folders, Pages files, entire Drives and even binary records from an installable program, migrated to an interoperable system.

Products

We also build our own products.

About

We are physicians and engineers. We build what we wish we'd had in the consulting room.

01
On one team
We know the consulting room, the night shift and the front office because we've worked there.
02
Specialized in Mexico
NOM-004, NOM-024, data protection and Ministry of Health catalogs, built in from day one.
03
Production experience
We run our own clinical workstation. We know what it takes to keep it running.
04
Privacy by design
Secure servers and zero data retention AI models: the AI provider doesn't keep your patients' data.

Founding team

A physician and an engineer. The same combination we bring to every institution.

“I'm not looking for technology that replaces people, but technology that amplifies what they can do.”
— Dr. Diego Lerma
DL

Dr. Diego Lerma

Co-founder and director

Physician from the University of Guanajuato, with a bachelor's in Software Development and a master's in Artificial Intelligence. Creator of Astrocyte; teaches at the University of Guanajuato and in FUNSALUD's AI in Healthcare program.

  • Medicine
  • Software development
  • Artificial intelligence
  • Education

Published in Springer Nature (2025) on AI adoption in Mexico, and at the ICLR 2026 MemAgents workshop.

doctordiegolerma.com
BA

Brandon Agüero

Co-founder · Engineering

Builds robust, interoperable, secure and scalable clinical systems. An expert in databases and in building high-level platforms, designing the architecture and backend of everything we build.

  • Databases
  • Backend
  • Architecture
  • Interoperability

Frequently asked questions

Something else on your mind? Write to us.

What kind of institutions do you work with?

Clinics, medical practices, hospitals and other healthcare institutions, public or private, of any size.

Do we have to replace our system?

Not necessarily. We can integrate with what you already use or, if it suits you, roll out Astrocyte across the hospital.

Where is patient data kept?

On secure cloud servers, processed with zero data retention (ZDR) AI models: the AI provider doesn't keep the information. We design every solution under NOM-004, NOM-024 and Mexican data protection law, and agree with you on who has access.

How much does it cost and how long does it take?

Every engagement is quoted according to its scope. If you want to see it working before deciding, we can prepare a demo in about a month.

Does the AI decide for the physician?

No. The AI structures and codes what the physician dictates or records, and flags data for review; the physician confirms everything before it enters the record.

Do I need to use your products to work with you?

No. Our services are independent from our products: we recommend whatever works best for your institution.

What do you want to solve at your institution? Tell us and we'll set up a call.