Six production endpoints — triage, lab interpretation, prescription OCR and clinical voice. One API key. PII redacted before any model sees it. Hosted in India.
# One request. Decision-support in ~1s. curl /api/v1/ai/triage \ -H "Authorization: Bearer ekai_live_9f2c…" \ -H "Content-Type: application/json" \ -d '{"issue":"chest pain", "duration":"2 days", "history":"hypertension"}'
{
"data": {
"urgency": "high",
"redFlags": ["cardiac risk", "refer ER"],
"advice": "Prioritise ECG + troponin…",
"disclaimer": "Decision-support draft."
},
"usage": { "cost": 1, "tokensRemaining": 4832 }
}
Every call is authenticated by API key, metered in tokens, and returns a structured JSON draft with a disclaimer. Bring your own UI.
No SDK lock-in — it's plain HTTPS + JSON. Works from any language, any stack, your servers or ours.
Provision an ekai_live_ key from your dashboard. Rotate or revoke anytime.
Send JSON over HTTPS. We redact PII, run the model, and return a structured draft.
Each response reports the token cost and balance. Overage and renewals follow your plan.
curl /api/v1/ai/summarize \ -H "Authorization: Bearer $EKAI_KEY" \ -H "Content-Type: application/json" \ -d '{"text":"58M, T2DM, presents with…"}'
const res = await fetch( "/api/v1/ai/summarize", { method: "POST", headers: { Authorization: `Bearer ${process.env.EKAI_KEY}`, "Content-Type": "application/json" }, body: JSON.stringify({ text: note }) }); const { data, usage } = await res.json();
import requests, os r = requests.post( "/api/v1/ai/summarize", headers={"Authorization": f"Bearer {os.environ['EKAI_KEY']}"}, json={"text": note}, timeout=60, ) data, usage = r.json()["data"], r.json()["usage"]
Personal & health identifiers are stripped before any model call — the LLM only ever sees de-identified clinical content.
Keys are stored only as SHA-256 hashes. Scoped, rotatable, instantly revocable.
Processed and stored in India — aligned with DPDP expectations.
Every token is reserved then settled; failed calls are auto-refunded, never charged.
Throttling and lockout per key stop abuse and protect your quota.
Every call is logged with key, endpoint and cost for reconciliation & review.
Outputs are always framed as drafts for a licensed clinician — never final advice.
Each plan includes a monthly token allowance that resets on renewal. Upgrade, downgrade or buy top-ups anytime. Prices in ₹, billed monthly.
Tokens are the unit of metering (1 token ≈ one triage/summarize call; OCR & voice cost more). Unused tokens don't roll over; top-up bundles available on every plan.
Create an account, generate a key, and make your first call in minutes. The first 1,000 tokens are on us.