Sequences a person's DNA, stores it securely, and turns it into clear, safe insights your app can display — as ready-to-share evidence summaries or plain-English answers.
Who it's for
Product teams building nutrition, longevity, recovery, or clinical-adjacent apps. No genomics background required — the API hands you finished insights, not raw biology.
What you can build
Personalized insight cards, medication-safety dashboards for clinicians, longitudinal health profiles, and bounded genome Q&A.
Under the hood, the same API orders 30x sequencing, delivers .genome/VCF/FASTQ files, hosts a genome runtime, builds genetic insight panels, and answers bounded natural-language genome questions.
How It Works
A genome moves through a few stages: you order sequencing, we deliver the files, we turn them into structured insights, and your app shows them safely. The names below show up throughout these docs.
The pieces, in plain words
How the runtime, interpretation artifact, panel, answer, and optional context artifact fit together.
Concept
In plain words
Order
The purchase. We ship a kit, sequence the returned sample, and deliver the files.
Delivery files
The raw genome data files: .genome, VCF, and FASTQ.
Hosted genome (runtime)
The genome kept online so you can keep asking it questions over time.
Interpretation artifact
The full analysis of the genome — large, detailed, and machine-readable. The source of truth.
Panel
The product-ready interpretation layer: user findings, genotype or score, meaning, evidence, citations, limitations, and safety boundary.
Answer
A plain-English reply to one question, with safety limits built in.
Context
Advanced: one topic of the genome as structured JSON for your own code or AI to read, not for direct display.
Order lifecycle
1created
2kit_shipped
3kit_delivered
4sample_registered
5sample_received
6lab_processing
7sequencing
8bioinformatics_processing
9completed
Runtime pipeline
1Order
2Delivery files
3Hosted genome
4Interpretation artifact
5Panel
6Answer
7Context (advanced)
Delivery files are the raw genome (the product of sequencing or conversion). A hosted genome runtime is an optional, durable copy you can keep asking over time — poll for updated interpretations, build panels, and ask bounded questions.
Five-Minute Quickstart
Start with our test genome — a fixed, synthetic genome that always returns the same result. You get real API responses without using anyone's real DNA, and no order is required.
TypeScript SDKnpm install @genomecomputer/sdk
Python SDKpip install genomecomputer
Use staging keygc_test_xxx
Build a panel from the seeded staging genome
export GC_API_KEY="gc_test_xxx"curlhttps://staging.api.genome.computer/v1/test/genomes \-H"Authorization: Bearer $GC_API_KEY"curlhttps://staging.api.genome.computer/v1/panels \-H"Authorization: Bearer $GC_API_KEY"curlhttps://staging.api.genome.computer/v1/genomes/gen_test_nutrition_001/panels/nutrition_micronutrients/build \-X POST \-H"Authorization: Bearer $GC_API_KEY"curlhttps://staging.api.genome.computer/v1/genomes/gen_test_nutrition_001/ask \-X POST \-H"Authorization: Bearer $GC_API_KEY" \-H"Content-Type: application/json" \-d '{"question": "Summarize this genome's nutrition panel in plain English.","response_length": "medium" }'
The same API surface works in production. Swap to https://api.genome.computer and a gc_live_ key only after contract, consent, billing, and pricing setup are complete.
End-to-End Guides
Recipes for the things teams build most often. Find the one closest to your product — nutrition insights, a pharmacogenomics dashboard, a longevity profile, or recovery personalization — and follow it end to end: the goal, the exact API calls, what comes back, what to show users, and a real response to copy from.
The goal
A nutrition feature
Show users a friendly profile of their diet-related genetics — no medical claims.
Steps
POST /v1/orders (or reuse a hosted genome)
GET /v1/genomes/{genome_id}/interpretations/latest
POST /v1/genomes/{genome_id}/panels/nutrition_micronutrients/build
POST /v1/genomes/{genome_id}/ask — optional, plain-English explanation
What you get back
A panel with a plain-language summary and a list of sections. Each section holds the user's actual finding, an evidence level, citations, and limitations.
What your app shows
Render each section as an insight card (title, finding, evidence level, citations). Skip raw variant tables unless your product is built for experts.
Real example
POST /v1/genomes/{genome_id}/panels/nutrition_micronutrients/build
{"panel_id":"nutrition_micronutrients","summary":"Folate processing, caffeine response, and lactose persistence for this genome.","sections":[{"title":"Folate & methylation","user_findings":"MTHFR C677T (rs1801133): one copy, moderately reduced enzyme activity.","evidence_level":"moderate","citations":["ClinVar"],"limitations":"Not a diagnosis. Does not measure folate levels."}]}
The goal
A PGx dashboard for a clinic
Give clinicians medication-response markers, ready for review.
Steps
POST /v1/orders with retention=hosted_runtime
Wait for the genome.created webhook
GET /v1/genomes/{genome_id}/interpretations/latest
POST /v1/genomes/{genome_id}/panels/pharmacogenomics_safety/build
Subscribe to interpretation.updated webhooks
What you get back
A pharmacogenomics panel that groups the user's markers by gene and drug-response area, each with an evidence level and a clinical boundary.
What your app shows
A clinician dashboard: gene rows with evidence badges and a clear "needs review" state shown before any medication guidance.
Real example
POST /v1/genomes/{genome_id}/panels/pharmacogenomics_safety/build
{"panel_id":"pharmacogenomics_safety","sections":[{"gene":"CYP2C19","drug_response_area":"Clopidogrel, some SSRIs","user_findings":"*1/*2 — intermediate metabolizer.","evidence_level":"high","clinical_boundary":"Clinician review required before any prescribing change."}]}
The goal
A longevity profile
Build a profile that gets better over time as the science does.
Steps
POST /v1/orders with retention=hosted_runtime
GET /v1/genomes/{genome_id}
POST /v1/genomes/{genome_id}/panels/longevity_healthy_aging/build
POST /v1/genomes/{genome_id}/panels/cardiovascular_safety/build
GET /v1/usage
What you get back
Runtime and panel records with stable IDs, the current interpretation version, and generated-at timestamps.
What your app shows
A profile that shows each panel's interpretation version, so users can see exactly when the evidence behind it changed.
Real example
POST /v1/genomes/{genome_id}/panels/longevity_healthy_aging/build
{"panel_result_id":"pan_l0ng1","panel_id":"longevity_healthy_aging","interpretation_version":"2026.05.1","generated_at":"2026-05-21T10:24:13Z","summary":"Cardiometabolic, inflammation, and lipid-transport context relevant to healthy aging."}
The goal
Recovery personalization
Add genetic context for inflammation, tissue repair, and recovery — while keeping clinical decisions on your side.
Steps
POST /v1/orders (or reuse a hosted genome)
GET /v1/genomes/{genome_id}/interpretations/latest
POST /v1/genomes/{genome_id}/panels/inflammation_recovery/build
POST /v1/genomes/{genome_id}/panels/connective_tissue_repair/build
POST /v1/genomes/{genome_id}/ask for a plain-English explanation
What you get back
Panels return inherited pathway evidence that can sit alongside your own labs and intake data.
What your app shows
Use the Q&A text as explanation, and keep treatment decisions inside your own clinician-reviewed workflow. Never present output as a treatment recommendation.
Real example
POST /v1/genomes/{genome_id}/ask
{"answer":"This genome shows inherited context in inflammation and tissue-repair pathways. Use it as background alongside labs and symptoms — it is not a treatment plan.","safety":{"clinical_advice":false,"requires_clinician_review":true}}
Synthetic genomic data only. Do not upload real genomic data or customer production data.
Production
Base URL
https://api.genome.computer
Key prefix
gc_live_
Purpose
Live sequencing orders, production genomes, production billing, partner applications, and real customer workflows.
Data policy
Real genomic data permitted only for approved partners under signed commercial, consent, and data-processing agreements.
Authentication and Key Management
API keys are environment-scoped bearer tokens. Treat production keys as live financial and genomic-data credentials.
Area
Behavior
gc_test_
Staging/sandbox only. Synthetic genomes, mock order advancement, webhook testing.
gc_live_
Production only. Real orders, real genomes, live billing, signed contract required.
Scopes
Keys can be limited by endpoint family, for example orders:read or webhooks:write.
Rotation
Create a new key, deploy it, confirm traffic, then revoke the old key.
Leaked key
Revoke immediately and review /v1/usage plus recent order events for unexpected activity.
Rotation pattern
# 1. Mint a new key in the dashboard/APIgc_live_new=gc_live_xxx# 2. Deploy your app with the new keyGC_API_KEY=$gc_live_new# 3. Revoke the old key after traffic confirmsDELETE /v1/api-keys/{api_key_id}
Sequencing Orders
Tier 1 orders deliver tracking, activation details, and secure file links for the deliverables requested by the client.
Create a sequencing order
The customer receives a kit activation identifier that can be used at sample.bio. Staging emulates the same flow without SLP or a real kit.
You decide at order time whether a genome is delivery-only or hosted for ongoing access.
Delete within 30 days
Use delete_after_30_days when the customer only needs secure delivery files. No hosted runtime fee is metered.
Hosted runtime
Use hosted_runtime to keep the genome addressable through /v1/genomes, interpretation, panel, context, and Q&A endpoints.
GET /v1/genomes/{genome_id} returns runtime metadata, not the genomic file itself. Use the bundle/files endpoints for signed downloads.
Genetic Insight Panels
Panels are the main way to turn a hosted genome into ready-to-show results. Each section gives the user's actual finding, what it means, how strong the evidence is, plus citations, limitations, and the clinical boundary.
What a panel adds
A panel is the product-ready interpretation layer. It runs a curated use-case definition against the hosted genome and returns the user's relevant variants, PGx markers, scores, interpretation text, evidence level, citations, and safety boundaries.
What it returns
Render sections as evidence summaries: the user's genotype or score, the panel-specific meaning, evidence basis, citations, limitations, and a clear boundary against diagnosis or treatment guidance.
Panel build response
{"panel_result_id":"pan_9y2m1q","panel_id":"nutrition_micronutrients","genome_id":"gen_test_nutrition_001","interpretation_version":"2026.05.1","summary":"This panel found user-specific nutrition and metabolic context with moderate evidence. It is suitable for an evidence summary card, not diagnosis or treatment advice.","sections":[{"title":"Metabolic Context","summary":"1 user-specific finding relevant to glucose handling.","interpretation":"For this user, TCF7L2 rs7903146 CT is relevant to glucose handling and diet-adjacent metabolic context.","evidence_level":"moderate","user_findings":[{"gene":"TCF7L2","variant_id":"rs7903146","user_result":"CT","interpretation":"The user carries one T allele at rs7903146, a common variant associated with glucose and type 2 diabetes risk in population studies.","relevance":"Useful as inherited metabolic context alongside diet, labs, history, and clinician review.","evidence_basis":"Moderate evidence based on GWAS Catalog and gnomAD."}]}],"citations":[{"source":"GWAS Catalog","url":"https://www.ebi.ac.uk/gwas/","accessed_at":"2026-05-20"},{"source":"gnomAD","url":"https://gnomad.broadinstitute.org/","accessed_at":"2026-05-20"}],"limitations":["DNA does not measure current glucose, insulin sensitivity, diet, or symptoms.","This panel does not diagnose diabetes or recommend a nutrition protocol."],"clinical_boundary":"Use as inherited context for education and clinician review. Do not present as diagnosis, treatment, or dosing guidance."}
Available insight panels
We add new panels continuously and build them hand-in-hand with partners. Want one tailored to your product? Email api@genome.computer and we'll design a custom insight panel for your use case.
Help a clinician-facing app explain inherited pathway context before peptide decisions.
MH
Metabolic Health
metabolic_health
Glucose handling, insulin signaling, lipid metabolism, body composition, and energy balance.
Power metabolic profile cards for nutrition and longevity products.
IR
Inflammation and Recovery
inflammation_recovery
Inflammatory signaling, oxidative stress, immune tone, endurance, and recovery-adjacent traits.
Support recovery, training, and protocol-personalization workflows.
CT
Connective Tissue Repair
connective_tissue_repair
Collagen, extracellular matrix, tendon, ligament, and wound-repair pathway context.
Surface inherited tissue repair context without making treatment claims.
CV
Cardiovascular Safety
cardiovascular_safety
Lipid transport, cardiometabolic context, medication-response markers, and safety signals.
Add safety context before apps discuss intensive interventions.
HE
Hormone and Endocrine
hormone_endocrine
Hormone-binding, endocrine signaling, thyroid-adjacent, and stress-axis genetic context.
Show inherited endocrine pathway context alongside labs and symptoms.
PGx
Pharmacogenomics Safety
pharmacogenomics_safety
CYP, transporter, guideline-linked, and medication safety markers.
Build medication response dashboards and clinician review queues.
NM
Nutrition and Micronutrients
nutrition_micronutrients
Folate, methylation, caffeine response, lactose persistence, and diet-adjacent traits.
Create nutrition intelligence cards with evidence boundaries.
LA
Longevity and Healthy Aging
longevity_healthy_aging
Cardiometabolic, inflammation, recovery, lipid transport, and age-related pathway context.
Combine multiple panels into a longitudinal healthy-aging profile.
Genome Q&A
Ask a genome a question in plain English and get a plain-English answer back. You send the question; we handle the AI and the safety rules behind the scenes.
Questions your users can ask
Each question is a single POST /askcall, so you can drop an “ask my genome” box or a full chat experience straight into your app. We handle retrieval, the model, and the safety rules.
How does my body handle caffeine?
Am I likely to be lactose intolerant?
Are there common medications I should be careful with?
What does my DNA say about recovery after exercise?
Do I have any nutrition traits worth knowing about?
Ask request
POST /v1/genomes/{genome_id}/ask
{"question":"How should my app explain peptide-related genetic context?","response_length":"medium"}
Natural-language answer
Response
{"answer_id":"ans_4az91r","genome_id":"gen_test_nutrition_001","response_length":"medium","answer":"This genome can support peptide-related decision making indirectly. The relevant genetic context is around metabolic handling, inflammation and recovery pathways, connective tissue repair, cardiovascular safety, endocrine signaling, and medication response. It should be presented as inherited pathway context for clinician review, not as a peptide prescription, deficiency assessment, or yes/no treatment recommendation.","citations":[{"source":"interpretation_artifact","version":"2026.05.1"}],"safety":{"clinical_advice":false,"requires_clinician_review":true},"usage":{"input_tokens":2190,"output_tokens":142,"model_name":"gpt-5.5"}}
The model should not provide clinical advice, dosing, prescriptions, or diagnosis. If a question asks for treatment guidance, the API responds with a boundary-setting answer that points back to clinician review.
Unsafe clinical request response
{"answer_id":"ans_refused_1","answer":"I cannot recommend a peptide, dosage, or clinical protocol from genomic data. The genome can provide inherited pathway context that a licensed clinician may consider alongside labs, symptoms, medical history, and medications.","safety":{"clinical_advice":false,"refusal_reason":"treatment_recommendation"}}
Context Builder (advanced)
For teams building their own logic or AI on top of the genome. context/build returns one topic as structured JSON that is always identical for the same input — easy to cache, test, and feed into your own model as grounding.
Not sure if you need this? You probably don't. Panels give you finished, ready-to-show results, and Q&A gives plain-English answers. Reach for context only when you want raw, reproducible genome data to feed into your own systems or AI.
{"context_id":"ctx_7v1a9k","genome_id":"gen_test_nutrition_001","domain":"nutrition","detail_level":"standard","summary":"Nutrition context includes folate handling, caffeine response, lactose persistence, lipid metabolism, and selected cardiometabolic markers.","structured_findings":[{"id":"nutr_folate_mthfr","title":"Folate and methylation context","evidence_level":"moderate","genes":["MTHFR","MTR","MTRR"],"client_display":"Show as inherited folate-pathway context, not as a diagnosis or supplement instruction."}],"citations":[{"source":"ClinVar","accessed_at":"2026-05-20","url":"https://www.ncbi.nlm.nih.gov/clinvar/"}],"interpretation_version":"2026.05.1"}
API Reference
The complete reference now has its own route at /reference. This summary keeps the everyday endpoints close to the guides.
Orders and delivery files
POST/v1/orders
Create a 30x sequencing order.
GET/v1/orders
List recent orders.
GET/v1/orders/{order_id}
Read order status and activation details.
GET/v1/orders/{order_id}/events
Read order lifecycle events.
GET/v1/orders/{order_id}/shipment
Read shipment carrier and tracking data.
GET/v1/orders/{order_id}/files
Return signed links for completed deliverables.
Uploads, conversions, and genomes
POST/v1/uploads/vcf
Upload an existing VCF.
POST/v1/conversions
Convert an uploaded VCF into a .genome bundle.
GET/v1/genomes
List hosted genomes.
GET/v1/genomes/{genome_id}
Read hosted genome runtime metadata.
GET/v1/genomes/{genome_id}/bundle
Download a signed .genome bundle link.
PATCH/v1/genomes/{genome_id}/retention
Change whether the genome is hosted or deleted.
Interpretations, panels, Q&A, and advanced context
All API errors use a consistent error object. Treat 400-series errors as request or configuration problems unless explicitly marked retryable.
HTTP
Code
Meaning
Client action
400
bad_request
Invalid JSON, missing required fields, unsupported status, or validation error.
Fix request and retry.
401
unauthorized
Missing, invalid, revoked, or environment-mismatched API key.
Do not retry until credentials are fixed.
403
forbidden
API key does not have the required scope.
Ask for a scoped key or change endpoint.
404
not_found
Order, genome, panel, webhook, or route does not exist for this organisation.
Check ID and environment.
409
conflict
Resource is in a state that cannot perform the requested action.
Refresh state before retrying.
429
rate_limit_exceeded
Default API limit is 600 requests per minute per key.
Back off and retry.
500
internal_server_error
Unexpected server failure.
Retry with backoff; contact support if repeated.
Standard shape
Error
{ "error": { "code": "not_found", "message": "Genome not found." }}
Retryable not-ready shape
Retry later
{ "error": { "code": "genome_not_ready", "message": "Genome runtime is still being created. Retry after the genome.created webhook fires.", "retryable": true }}
Default rate limit is 600 requests per minute per key. List endpoints use a limit query parameter today; cursor pagination will be added before the lists become large.
Endpoint
Pagination
Notes
GET /v1/orders
limit 1-100
Newest first. Cursor pagination will be added before high-volume partner rollout.
GET /v1/genomes
limit 1-100
Newest first, organisation-scoped.
GET /v1/usage
limit 1-100, event_type
Newest first with optional event type filtering.
Polling
15-60 seconds
Prefer webhooks for lifecycle changes; avoid aggressive order polling.
Webhooks
Use webhooks for order and runtime lifecycle events instead of polling aggressively.
Immediately after an API sequencing order is accepted.
order
order.kit_shipped
When fulfillment records outbound kit shipment.
order, shipment
order.kit_delivered
When carrier tracking reports kit delivery.
order, shipment
order.sample_returned
When return shipment is registered or received by carrier.
order, shipment
order.sample_received
When the lab receives the sample.
order
order.lab_processing
When lab processing begins.
order
order.completed
When sequencing/conversion deliverables are ready.
order, genome_id, files_url
conversion.completed
When an uploaded VCF has produced a .genome bundle.
conversion, genome_id
genome.created
When a hosted genome runtime is created.
genome
interpretation.updated
When a new interpretation version is attached.
genome_id, interpretation
context.generated
When a context artifact is built successfully.
context
panel.generated
When a genetic insight panel is built successfully.
panel
Deliveries retry with exponential backoff from roughly one minute to one hour. Return any 2xx status to acknowledge receipt. Non-2xx, timeout, DNS, and TLS failures remain pending and are retried.
The API is versioned at the path level. Partners should tolerate additive fields and pin integrations to /v1.
Policy
Details
Current API
/v1
Breaking changes
Introduced under a new version path when possible, with partner notice before migration.
Additive changes
New fields may appear in existing responses. Clients should ignore unknown fields.
Deprecation
Production partners receive a migration window before an endpoint or field is removed.
Billing
Staging records metered events for testing. Production usage is billable under the organisation pricing configured during onboarding.
30x WGS$699/order
.genome conversion$99
Hosted runtime$3/mo
Context/panels/Q&A$0.02
Contexts, panels, and Q&A are billed per successful generation, not per token. Internally we still record model name, input tokens, output tokens, cost, and latency for margin monitoring.
Glossary
Plain-English definitions of the genomics and clinical terms used across these docs.
Term
Meaning
30x WGS
Whole genome sequencing where each base is read about 30 times on average.
.genome
The canonical genome.computer bundle: structured genomic data, manifest, provenance, and interpretation-ready indexes.
VCF
Variant Call Format. A file listing genetic variants compared with a reference genome.
FASTQ
Raw sequencing reads and quality scores. Useful for reprocessing, but much heavier than a VCF or .genome bundle.
PGx
Pharmacogenomics: inherited variation that may affect medication metabolism, response, or safety.
Interpretation record
A lightweight database record describing the current interpretation version and status.
Interpretation artifact
The full structured interpretation output used by panels, Q&A, and advanced context artifacts.
Panel
The main structured interpretation product: user findings, actual result or score, meaning, evidence strength, citations, limitations, and safety boundary.
Context
Advanced: a reproducible, structured JSON summary of one genome topic. Built for your own systems or AI, not for direct display.