It's time to build bio: introducing the Genome API

One API between your product and the genome

Building on genomic data should not mean scraping consumer reports, integrating closed interpretation SDKs, or standing up your own lab.

The Genome API gives your product a direct path to the underlying genome. Programmatically order 3x/30x whole-genome sequencing or GSA genotyping, keep a hosted genome runtime, build insight panels, and ask bounded genome questions. We handle kit fulfillment, lab processing, pipeline execution, and result delivery.

The same API also accepts existing VCFs. If your user already has data from a legacy provider, clinical run, or research dataset, you can convert it into the same .genome output surface without sequencing them again.

What you get

• Programmatic sequencing and fulfillment

  Create an order, send a kit to your customer, track the sample through the lab, and receive results when processing completes.

  curl https://api.genome.computer/v1/orders \
    -H "Authorization: Bearer $GENOME_API_KEY" \
    -H "Idempotency-Key: order-user-123-2026-05-21" \
    -H "Content-Type: application/json" \
    -d '{
      "type": "wgs_3x",
      "client_name": "Acme Health",
      "user_reference": "user_123",
      "customer": {
        "name": "Avery Chen",
        "email": "avery@example.com"
      },
      "shipping_address": {
        "line1": "535 Mission St",
        "city": "San Francisco",
        "region": "CA",
        "postal_code": "94105",
        "country": "US"
      }
    }'

• A .genome/1.0 bundle

  Every result is delivered as a structured, AI-readable genome bundle: typed, deterministic, queryable, and pinned to the pipeline version, reference build, and source snapshots used to generate it.

  {
    "schema_version": "1.0.0",
    "pipeline_version": "genome-api-2026.04",
    "genome_build": "GRCh38",
    "generated_at": "2026-04-20T22:44:34Z",
    "annotation_sources": {
      "vep": "111",
      "clinvar": "2026-04",
      "gnomad": "4.1",
      "acmg_sf_list": "v3.2",
      "cpic_guidelines": "2026-03"
    },
    "consensus_definitions": {
      "is_actionable": {
        "rule": "ACMG SF gene + pathogenic ClinVar + >=2 review stars",
        "gene_list_sha256": "..."
      }
    },
    "files": {
      "variants.parquet": {
        "sha256": "...",
        "rows": 4823921
      }
    }
  }

• Interpretation and agents

  Once a genome is hosted, your product can reason over it without building interpretation in-house. Build curated insight panels — nutrition, pharmacogenomics, metabolic, longevity, and more — that return findings with evidence levels, citations, and limitations. Ask bounded natural-language questions and get plain-English answers with provenance. Agents can query the genome directly through our MCP server for gene-level pharmacogenomic and variant calls.

  Every panel and answer is bounded by design: no diagnosis, no dosing, no prescriptions. Results are framed for clinician review and carry explicit safety metadata.

  curl https://api.genome.computer/v1/genomes/gen_123/panels/pharmacogenomics_safety/build \
    -X POST \
    -H "Authorization: Bearer $GENOME_API_KEY"

• gVCF input compatibility

  Existing gVCFs can be uploaded and converted for downstream workflows. .genome is the software-native delivery layer.

• Bring your own gVCF

  Convert existing genomic files into the same downstream format as a fresh sequencing order, typically within 24 hours. One output surface, whether the genome starts with a new kit or an uploaded file.

• Accredited lab infrastructure

  Sequencing runs through CLIA- and CAP-accredited laboratory infrastructure in the United States, with HIPAA-aligned controls and reproducible bundle generation.

• Built on a Public Benefit Corporation

  Genome Computer operates under Genetic Superintelligence Company, a Delaware Public Benefit Corporation. Our charter commits us to keep identifiable individual genetic data under individual control and out of commercial sale or licensing. That commitment is part of the corporate structure the company is governed by, not just the current version of a terms page.

Why .genome matters

gVCF is built for genomic analysis pipelines. .genome is built for software readers.

A .genome/1.0 bundle separates variant data, interpretation, and rules into explicit, versioned structures. Agents do not have to infer what a field means, guess which guideline version applied, or reconstruct clinical significance from loosely encoded annotations. The schema carries the meaning with the data.

That makes genomic reasoning faster, more auditable, and less fragile: applications can query the genome directly, cite provenance, and compare future reinterpretations against the exact snapshot a user received.

Prefer email? api@genome.computer