The Genomic Intelligence API is live

Our sequence-to-function models are now available programmatically. Today we are launching the Genomic Intelligence API — a typed /v1 REST API and an MCP server that expose the same models powering the DNA Analysis Platform, so you can run predictions from your own code, pipelines, and AI agents.

Full documentation lives at docs.genomicintelligence.ai.

Six tasks, one contract

The API serves six DNA sequence analysis tasks, each at a single endpoint — POST /v1/tasks/{task}/predict:

• Promoter — classify whether a sequence contains a promoter region.
• Splice — annotate donor and acceptor splice sites across long sequences.
• Enhancer — score developmental and housekeeping enhancer activity.
• Chromatin — predict 919 chromatin features (DNase, CTCF, histone marks, …) as grouped tracks.
• Expression — predict gene expression from sequence plus an experimental description, in log(TPM+1).
• Annotation — find genes by detecting TSS and PolyA signals, emitting transcript intervals.

Every endpoint shares one shape, so learning one task teaches you all six:

• API key on every /v1/* route, sent as Authorization: Bearer gi_….
• One success envelope — {data, meta} on every 2xx and 202. data is task-specific; meta is uniform across tasks.
• One error envelope — {error: {code, message, request_id, details?}}. Switch on error.code, never the message.
• Sync by default, async on request — add Prefer: respond-async for long inputs: the call returns 202 with a job_id you poll for.

The live OpenAPI schema (browse it in ReDoc) is the machine-readable source of truth for every field.

Making a first call

Ask us for a key to get a gi_… token. The same key works for the REST API and the MCP server. Verify it in one call:

curl -sS https://api.genomicintelligence.ai/health
# {"status":"healthy","version":"YYYY.MM.DD.iter (commit)"}

From there, drop the single-file Python client from the integration kit into your project and call a task:

import os
from gi_client import Client

client = Client(api_key=os.environ["GI_API_KEY"])

body = client.predict("promoter", sequence="ACGT" * 500, sequence_name="demo")
print(body["meta"]["model"], body["meta"]["inference_time_ms"], "ms")
print(body["data"]["regions"])

The kit ships a runnable quickstart that hits every task on real bundled sequences, plus recipes for the common patterns — gene-list sweeps, async polling, rate-limit-aware retries, and typed error handling. The REST API guide walks the whole contract end to end.

Use it from an AI agent

The same six tasks are exposed as tools over an MCP server, so you can ask questions over DNA in Claude, ChatGPT, or your own agent and let it call the models directly — “Fetch a 100 kb window around human TP53 and scan it for promoter regions” or “Predict HBB expression in K562 cells.” You can start with no key and no install through the hosted GI connector, then move to a local server for production work. See the MCP guide to get connected.

What’s next

The API and MCP server reach every model we serve, and they grow as the platform does. New tasks and models land in the same {data, meta} contract, so integrations pick them up without rework. Grab a key, read the docs, and start scripting against your own sequences.

These are computational predictions intended for research and development, not clinical or diagnostic decisions. For high-volume or latency-sensitive use, contact us to provision a production-shaped deployment.