Loading KohiShifa…
Loading KohiShifa…
AI Wellness CompanionComing soon
An optional, consent-first companion for routines, reflection prompts and planning. It isn’t available on this deployment yet — here’s exactly what it will and won’t do, and how it’s designed to be safe.
Status on this site
AI companionComing soon
Not available on this deployment yet. The integration is built and will be switched on once it can run safely behind sign-in.
Accounts & sign-inAvailable
Needed to use the companion.
Everything elseWorks without AI
Habits, hydration, sleep routine, reflection, mindfulness, activity, goals and insights.
Not available on this deployment yet
What it can help with
The companion is designed for the practical side of looking after yourself: ideas, routines, plans and gentle encouragement. These are the kinds of questions it’s built for.
Example questions you might ask — written by us, not answers from the AI.
“Help me turn “move more” into something I can do on weekdays.”
“What could a calmer morning look like if I only have 15 minutes?”
“How can I protect a lunchtime walk on busy days?”
“Talk me through a two-minute breathing pause.”
“Why do people suggest keeping a regular bedtime?”
“Give me three prompts for looking back on a hard week.”
“I’ve missed my habits all week. How do I restart gently?”
Firm boundaries
A wellness companion is not a clinician. These limits are built into how the companion is instructed and how its replies are checked. Ask for any of them and it’s designed to say so plainly and point you towards a qualified professional.
Diagnose conditions or interpret symptoms
Recommend, start, stop or adjust medication
Prescribe diets, meal plans or calorie targets
Design intense or high-risk exercise programmes
Claim to be a doctor, therapist, dietitian or any other professional
Replace care from a qualified professional
A preview
We wrote this example ourselves to show the tone we’re aiming for: practical, brief, and honest about limits. It is not a live AI-generated response.
Example reply · written by the KohiShifa team
A few small ideas that often help:
Your daily goal is yours to choose — if you have specific fluid needs or a health condition, a healthcare professional can tell you what’s right for you. Want help picking one “anchor” moment to start with?
Example reply · written by the KohiShifa team
Sure — try a 20-minute buffer: dim the lights, set out tomorrow’s things, then read something light. If you’d like, add “Wind-down at 10:30” as a weekday habit so it shows up in Today.
Safety by design
No safety system is perfect, so the companion is built with several layers rather than one.
Before your first message you’ll see what the companion is, what it isn’t, and what is sent to the configured AI provider. Nothing is sent until you agree.
Messages are screened for crisis language. If something is flagged, you’ll see crisis-support resources instead of an AI reply.
Replies are checked before they’re shown, to catch content that strays outside general wellness — like medication or diagnosis.
A daily message limit keeps the companion a thoughtful tool rather than an endless feed — and keeps it sustainable for us to run.
Each response is marked as AI-generated, so it’s always clear what came from AI and what came from you or from KohiShifa.
KohiShifa doesn’t send an AI reply. Instead it shows crisis-support details like these, straight away. Keyword screening can miss things, so please don’t rely on any app in an emergency.
Shown instead of an AI reply
If you’re thinking about harming yourself or feel unsafe, please contact someone who can help right now.
United States
988 Suicide & Crisis Lifeline
United Kingdom & Ireland
Samaritans
Emergency
Emergency services
Call 911 (US) or 999 (UK)
Elsewhere, please contact your local emergency number or a local crisis line.
Privacy
The companion is off until you choose to use it, and even then it only sees what you send.
Powered by the selected Gemini or Claude model
KohiShifa uses Google Gemini or Anthropic Claude through a server-side API, depending on the deployment. Neither provider endorses or reviews KohiShifa. Our Privacy Policy explains how messages are processed.
Your messages go to the configured Google Gemini or Anthropic Claude provider to generate a reply only when you send one. If you never open the companion, nothing is sent.
A short summary of your habits is included only if you switch it on for that message. It’s off by default.
Your reflection journal is never sent to the companion automatically. If you want to discuss something, you choose what to type.
KohiShifa doesn’t keep the content of your companion conversations on our servers.
Questions
Straight answers about availability, the model, boundaries and your data.
Coming soon
The companion isn’t available on this deployment yet. The demo works fully without AI — every tracker, insight and exercise — and early access is the best way to hear when it’s switched on.