Introducing Genomi · Genomi<br>Why nowThree problemsThe harnessTrustworthyPrivacyDynamicDocs Agent SetupMost people who have a DNA file do not have a genome. They have a zip file sitting somewhere, maybe a 23andMe export, an AncestryDNA text file, a VCF from a sequencing company, or a static report with bright charts and a lot of hedging.<br>The file is real. The biology is real. But the experience is strangely dead. You cannot ask it follow-up questions. You cannot ask why one source disagrees with another. You cannot ask whether a variant was measured, whether the region was missing, or whether a medication question needs CYP2D6 and CYP2C19 evidence before anyone should trust the answer.<br>So the most personal dataset you own becomes something you download once, skim once, and then forget. That feels wrong.<br>We built Genomi because we think the genome should become a living workspace for your AI agent: not a fortune teller, not a replacement for a clinician, and not a startup vault where your DNA goes in and investor updates come out. Genomi is a local, evidence-grounded harness that lets your agent work with your DNA carefully.<br>Why now<br>Personal genomics is becoming urgent because four forces are colliding at the same time.<br>Sequencing keeps getting cheaper<br>NHGRI's sequencing-cost dataset tracks the drop from roughly $95M per genome in 2001 to roughly $562 in 2021.1Production-scale platforms such as Illumina's NovaSeq X pushed the industry toward more than 20,000 genomes per year per instrument.2 Sequencing is moving from a rare medical event toward a personal data layer.<br>Genomics discovery is accelerating<br>ClinVar now reports more than 6.8 million submissions across more than 4.5 million variants.3 The NHGRI-EBI GWAS Catalog has reported more than 625,000 curated lead associations across more than 15,000 traits.4 Those numbers mean a DNA answer is increasingly a moving bundle of sources, assertions, confidence levels, ancestry caveats, clinical boundaries, and sometimes disagreement.<br>AI is entering that research loop too. AlphaFold made more than 200 million predicted protein structures available.5 AlphaMissense predicted effects for 71 million possible missense variants.6 AlphaGenome is applying AI to regulatory variant-effect prediction.7 Google Research has introduced AI co-scientist systems, and the FDA approved the first CRISPR/Cas9-based therapy for sickle cell disease in December 2023.89<br>Health agents need genetic context<br>People already ask AI about symptoms, medications, nutrition, sleep, exercise, lab results, family history, and whether some odd thing is worth worrying about. If a health agent does not understand your genes, it is missing one of the most personal layers of your biology. Genes are not destiny, but they are context.<br>Edge computing is becoming real<br>The local machine is turning into an AI runtime, not just a thin client for cloud calls. On June 1, 2026, The Guardian reported Nvidia's RTX Spark push for Windows PCs and laptops, describing chips designed to run AI agents locally rather than relying on cloud computing.10 That matters for genomics: the more useful edge AI becomes, the more reasonable it is to keep deeply personal files on your own machine and let an agent work through a local, auditable index.

The three problems<br>You should not have to hand your DNA to startups<br>DNA is not normal data. It identifies you, says things about your family, does not expire, and cannot be rotated like a password. The 23andMe bankruptcy made this less theoretical: in 2025, the FTC raised concerns about potential sale or transfer of 23andMe user data.11 People should be able to use their DNA without surrendering it.<br>No static report can keep up<br>A static DNA report starts aging the moment it is generated. A variant gets reclassified. A drug-gene guideline changes. A population-frequency source adds context. A phenotype-gene relationship gets stronger, weaker, or more complicated. The genome is a living evidence problem, and we keep packaging it like a one-time document.<br>General AI can sound right and still be wrong<br>A model can explain APOE, BRCA1, CYP2C19, or lactose intolerance without knowing your answer. For a real person, the questions are more concrete:<br>Do I have the relevant variant?<br>Was it measured, and was the region callable?<br>What genome build is this on?<br>What do ClinVar, gnomAD, CPIC, PharmGKB, PGxDB, or FDA sources say?<br>Is this clinical, preliminary, population-level, or not applicable?<br>That is not prompt work. That is tool work.

The harness<br>Genomi is an open source AI agent harness that turns your AI agent into a personal DNA expert. It gives your agent a local genomics workspace: a private index of your genome, evidence tools, public genetics source access, an investigation journal, and a dashboard.<br>The agent handles the conversation. Genomi handles the parts where guessing is not acceptable. It works through MCP-capable agent hosts, including Claude Code, Codex, OpenCode, OpenClaw, Hermes,...