Local-first · Private by design

A medical record workspace that never leaves your computer

OpenHealth brings every patient's records into one workspace and lets you ask grounded questions about them — powered by AI that runs entirely on your own machine. No cloud. No account.

View on GitHub See how it works

Free & open source · Runs on macOS & Windows · Powered by Ollama

localhost:3000
J. Rivera
M. Chen
A. Okafor
+ Add patient
What were the last lab results and any flagged values?
The most recent panel (Mar 14) shows HbA1c at 7.2% — slightly elevated from the prior 6.8%. Cholesterol and CBC are within normal range.
↳ Grounded in: labs_2026-03-14.pdf, visit_notes.pdf

Everything you need, nothing you don't trust

Built for clinicians and caregivers who want answers from their records without handing them to a third party.

Private by design

No account to create and no cloud service in the loop. Documents and details you add never leave the device.

Grounded answers

Every response is anchored to the actual patient record, with citations back to the source documents it used.

Local AI via Ollama

Answers come from medical-tuned models running on your own hardware. No internet connection required.

Per-patient workspaces

Organize records into separate patient profiles. Add age, conditions, and history to sharpen every answer.

Drop in any document

Upload PDFs and scans — built-in OCR reads even image-based records so nothing gets left behind.

Timeline view

See a patient's records and ingests laid out over time, so the full history is one glance away.

Up and running in four steps

Designed for non-technical users. The installer checks Docker, sets up Ollama, and pulls the AI models for you.

1

Install Docker

Download Docker Desktop for macOS or Windows and launch it.

2

Run the installer

One script sets up Ollama and pulls the required models for you.

3

Start the app

Run docker compose up from the project folder.

4

Open your browser

Go to localhost:3000 and create your first patient.

Your patient data stays on your machine — full stop

OpenHealth runs entirely on your computer. Records are stored and read locally, and full LLM payloads are redacted from logs by default to protect PHI.

0Cloud calls
0Accounts needed
100%On-device

Bring your records into focus

Free, open source, and private. Get started in minutes.

Clone the repo Read the docs

Requires 16 GB+ RAM · ~30 GB free disk · Docker Desktop & Ollama