How to read your 23andMe raw data privately, in your browser

If you have ever taken a 23andMe or AncestryDNA test, you already own a raw DNA file. You can read what is actually in it without handing your genome to another company’s server. Here is how.

What “raw data” actually is

Your raw file is a plain text export: roughly 600,000 to 700,000 markers (SNPs), each with an rsID, a chromosome position, and your genotype, like AG or TT. It is not a report. It is the input a report gets built from, and the company’s own health summaries only interpret a small slice of it. The rest sits unused until you decide to read it.

Download your file first

In your 23andMe account: Settings, then Download raw data, confirm your identity, and wait for the export. You get a .txt file, sometimes zipped. It does not expire, and it works with any tool built for the standard 23andMe or AncestryDNA format. Keep your own copy somewhere you control. It is the one thing you cannot get reissued if a company disappears.

The privacy question nobody asks: where does the file go?

Every interpretation tool falls into one of two camps. Upload-based tools take your entire raw file onto their servers, parse it there, and keep it under whatever retention policy they run. Local-parsing tools read the file in your browser’s memory and never send the file itself; at most they check the handful of specific marker positions a report needs.

The difference matters because raw DNA is closer to a fingerprint than a password. It cannot be rotated if it leaks, and it exposes your relatives too. The 23andMe bankruptcy, in which its genetic database became a sellable asset, is the clearest recent reminder that “where does my data sit, and who owns the company holding it” is not a hypothetical.

The 30-second test that works on any tool

Before you upload anything, open your browser’s developer tools (press F12) and go to the Network tab. Select your file in whatever tool you are evaluating. If you see a large outbound request roughly the size of your file, usually 15 to 30 MB, it is uploading. If your file size never appears in outbound traffic, it is being parsed locally. This check takes less time than reading a privacy policy, and it does not rely on trusting anyone’s marketing.

What you can actually learn

The highest-signal information in a raw file is pharmacogenomics: how your genotype affects your response to specific medications. CPIC, the Clinical Pharmacogenetics Implementation Consortium, publishes evidence-graded guidelines for drug and gene pairs, and your file already contains the genotypes to check most of them. This is genuinely useful context to bring to a prescriber, never something to act on alone. Trait predictions are a mixed bag: some rest on large replicated studies, many on a single small one. A tool worth using tells you which is which instead of presenting everything with the same confidence.

Where Verinome fits

Verinome parses your raw 23andMe or AncestryDNA file entirely in your browser. You can watch the Network tab and confirm the file never uploads. To build the full $49 report, your browser sends only about 200 specific marker readouts, a few kilobytes with no file and no name attached, and nothing is stored afterward. Every finding carries an A to D evidence grade, and the weak ones are labeled, not hidden. There is a free preview, no account and no card, that runs the same file-quality check and shows a starter set of graded results before you decide.

See exactly what leaves your browser on our privacy page, watch your file not upload, or compare the options side by side.

Educational, not medical advice. Pharmacogenomic and trait information should be discussed with a qualified clinician before any health decision.