How Ojava’s AI works

What Ojava Coach is

Ojava Coach is an AI assistant that helps you make sense of YOUR OWN health information. It reads the records you accept (labs, wearables, sleep logs, fitness data, medications, notes) and helps you understand patterns, prepare for visits, ask better questions, and think through next steps.

It is not a doctor, not a clinician, not a diagnostic tool, and not licensed medical software. It is an AI that is good at connecting dots across the health data you already have and helping you organize information before you talk to a real healthcare provider.

Grounded in real numbers, not guesses

Ojava Coach does not invent numbers. Behind the scenes, Ojava runs pure calculators to turn your raw records into finished values: your readiness score, sleep debt, biological-age inputs, fitness age, training load, marker status, and habit signals. These are deterministic, always the same result given the same input.

When Ojava Coach answers a question, it reads those finished, computed values, not your raw rows. If a number is missing, it says that plainly instead of guessing. If you ask about a biomarker and the value is not in your record, the Coach will tell you what value would help and what the next step is, but it will not invent one.

This matters because it keeps the AI from hallucinating health numbers. Your guidance is based on what you actually have, and the Coach can point exactly to what is missing to make answers more reliable.

What Ojava Coach will never do

There are hard boundaries between wellness coaching (which the AI can help with) and licensed medical practice (which it cannot). Ojava Coach will never:

• Diagnose a disease or name a condition you have.
• Prescribe a medication or supplement, or tell you the dose or timing.
• Order or request a lab test on your behalf.
• Declare that a medical result is safe or unsafe for you specifically.
• Claim to replace, substitute for, or be equivalent to a licensed clinician.
• Tell you to skip, delay, or ignore medical advice you received from a doctor.

These boundaries are not policy we enforce softly, they are hard limits built into how the AI works. It cannot be talked out of these boundaries, even if you ask directly.

Safety comes first, before anything else

Before Ojava Coach runs at all, a deterministic safety check scans your message for signs of a medical emergency or a mental-health crisis. If it detects self-harm language, suicidal thoughts, or an emergency symptom, it stops. It does not ask the AI for an answer. Instead, it gives you a blocking message directing you to call 911 or text 988 (the US Suicide and Crisis Lifeline), with clear action buttons right in the chat.

This check cannot be overridden by what you ask or how you phrase it. It is a deterministic rule, not the AI’s judgment.

Ojava also watches what you ask about, and if you ask for something outside health and wellness (code, trivia, legal advice, financial decisions), it gently points you back to what it is designed to help with instead of answering the off-topic question.

Privacy: your data stays yours

Ojava does not use your health data to train models by default. Your records are used to power the answers Ojava Coach gives you; they are not fed into AI model training, sold, or shared with third parties.

Sensitive health information (mental health, substance use, reproductive health, genetic data, HIV-related records) receives extra care. Ojava identifies sensitive records and handles them carefully, withholding them from certain AI outputs by default unless you explicitly consent.

You can export your data at any time, and you can delete your account and all your records. If you have questions about your privacy, read the full privacy policy at /docs/privacy.

Honest about limits and uncertainty

When Ojava Coach does not have enough information, or when it is uncertain, it says so. It separates what is grounded in your actual record from what would be a guess. It does not dress up uncertainty with false confidence or invented certainty scores.

If you ask what it knows about you, it will tell you exactly what is in your record. If you ask it to compare options and it does not have the data it needs, it will ask you what is missing instead of inventing details.

When the answer would benefit from a licensed professional, the Coach will tell you that too. It is not trying to be all things; it is trying to be honest about what it can and cannot do.

Educational, not clinical

Ojava Coach explains health concepts and helps you understand your own data. It is not clinical advice, not a treatment plan, not a diagnosis, not a substitute for seeing someone.

It can help you prepare for a visit (here are the questions to ask, here is the context to bring). It can help you understand a result your doctor already gave you (here is what that marker means, here is what changes over time). It can help you track patterns (your sleep looks different this week, your workouts cluster on these days). But it cannot decide what you should do medically.

When Ojava Coach drafts a packet for a clinician (like a visit summary or a list of questions), that packet is for your review and your eyes first. You decide whether to share it, edit it, or use it. Ojava does not send anything to a clinician on your behalf.

Your record is data, never instructions

Ojava treats your health records as information, never as commands. If a note or a field in an imported document says something that tries to change how the AI behaves, that cannot happen. Your data is context, not control. This boundary is built in.

Ask your clinician the question the Coach helped you draft

After most conversations, Ojava Coach can help you draft three clear questions to ask a clinician or therapist. Those questions live in a copyable text block so you can move them into your notes. This is the bridge between Ojava understanding and professional care.

The AI explains and prepares. Your clinician decides and acts. That boundary is where trust lives.