# one call, calibrated answer curl https://api.lexun.co.uk/v1/decide \ -H "Authorization: Bearer $LEXUN_KEY" \ -d '{ "question": "Should we open a second site?", "context": { "sector": "hospitality", "runway_months": 9 } }' # response { "odds": 0.63, "band": [0.51, 0.74], "stance": "TEST SMALL FIRST", "best_move": "Pilot 3 months before signing", "drivers": ["base_rate", "runway"], "decision_id": "dec_8k2m…" // report the outcome later → engine recalibrates }
Monte-Carlo simulation anchored to sourced base rates, with an explicit confidence band — never a bare, unaccountable number.
POST the real result back and the model recalibrates — your integration gets measurably sharper, and you can show users the accuracy record.
Every response carries its drivers and its uncertainty. When the engine doesn't know, it says so — that honesty is the product.
Send only the context you choose. No profiles built, no data resold — the same local-first ethic as the LEXUN app.
What teams build with it
Affordability and risk reads with a confidence band your compliance team can defend.
Role-fit and automation-exposure odds inside your existing workflow.
Calibrated scenario odds for underwriting and claims triage — scored against outcomes.
Give every recommendation an honest probability and a next move.
Buy-vs-rent and where-to-live reads embedded in your product.
Policy scenario odds with the auditable calibration record built in. See Business & Government →
We're onboarding a small first cohort of design partners — free during beta, usage-based pricing at GA. Tell us what you're building and we'll get you a key.
The API is in private beta; the endpoint shown illustrates the target interface and may evolve with design partners. LEXUN returns decision intelligence, not financial, legal or medical advice.