Set up Claude for reading lab results in 10 minutes

The most careful chat model on the market, configured for medical documents.

8 minutes start to finish

Claude is the model I reach for when the document is medical. It pushes back harder on overreach, cites more honestly, and is generally less seduced by patterns that aren't there.

Before you start

  • A free claude.ai account
  • Your most recent full blood panel (PDF or photo)
  • 5–10 minutes

The steps

  1. 01

    Create a Project

    Top-left sidebar → New Project. Name it "Personal Health". Projects keep documents and instructions scoped — far better than dumping everything into one chat.

  2. 02

    Add project instructions

    Paste: "You're helping me read my own health data. Cite primary literature when possible. Distinguish observational from interventional evidence. Never diagnose; flag what belongs with a clinician. If a value is outside reference range, explain the most common benign and concerning explanations — in that order."

  3. 03

    Upload your panel + the last one if you have it

    Trend matters more than any single number. Ask: "Compare these two panels. Which markers moved in a clinically meaningful direction and what's the most parsimonious explanation?"

  4. 04

    Save the project, return weekly

    The next time you have new labs or a question, open the same project. You're building one continuous case file Claude can reason against — not 30 disconnected chats.

Honest note

Claude is careful, not omniscient. If it tells you to talk to a clinician, do that — it's usually right.

Want the whole stack, not just one tool?

The free 10-Day Challenge wires these together. Or join the free 45-min live workshop and watch me build it end-to-end.

Want the whole method, not just one tool?

The free 10-day email challenge installs Research → Ledger → Protocol on whatever data you already collect.

More for people setting up their stack

Based on what you've been reading — always learning.

See all →