What Saathi Assist delivers.
What's working today, in user hands.
Live features people use in every session — not a roadmap.
Daily Check-Ins
A short, structured rhythm: mood, energy, sleep, and focus — plus a free-text moment. Up to two check-ins per 24 hours, with a 12-hour gap, so reflection stays meaningful instead of mechanical.
Oral Health Tracking
A dimension most wellness tools ignore. Saathi logs pain, bleeding gums, tobacco use, and quit intent — gently surfacing patterns that matter for whole-person health.
Assessments & Medical Reports
PHQ-9 and GAD-7 are administered, scored, and interpreted in-conversation. Uploaded medical documents are read by Saathi — extracted vitals, medications, diagnoses, and allergies are shown back to the user, who approves each update before it enters their profile.
Choose how you want to be met.
Text, voice, or a real-time 3D video avatar. The same companion, the same memory, the same safety.
Text chat
Quiet, on-screen conversation for moments when typing feels easier than speaking.
Natural voice
Talk to Saathi the way you would talk to a calm friend on a phone call.
Real-time 3D avatar
A face-to-face video presence for people who feel seen by being looked at gently.
Meeting people in the language they live in.
English, Hindi, and Arabic with full right-to-left layout — so multilingual workforces and citizens are not asked to translate themselves into care.
Every chat turn, monitored in real time.
Most wellness tools wait for a person in distress to self-identify. The Saathi safety orchestrator removes that burden — it watches every turn for escalating risk signals and responds proactively.
Signals of acute risk trigger an immediate, calm response in-conversation and surface verified crisis resources.
Example. Direct mentions of self-harm or harm to others.
Sustained distress is recognised across turns. Saathi slows the conversation, validates, and gently routes toward human support.
Example. Persistent hopelessness, severe sleep loss, escalating panic.
Patterns suggesting clinical follow-up are flagged with a soft nudge to a professional, with the user's history ready to share.
Example. Repeated high PHQ-9 / GAD-7 scores or persistent physical symptoms.
100% of chat turns are monitored. Every escalation is logged for clinical and governance review.
Validated screening with care, not content.
PHQ-9 and GAD-7 are administered, interpreted, and connected to next steps — so wellness has a measurable foundation, and referrals are credible.
A four-layer care platform.
Modular by design, governed by default.
Care companion layer
The conversation surface in text, voice, or 3D avatar across three languages.
Clinical intelligence layer
Validated screening (PHQ-9, GAD-7), interpretation, and longitudinal wellness scoring.
Safety orchestrator
Real-time monitoring of every chat turn for escalating risk signals, without the user having to ask.
Governance & integration layer
Identity, audit trails, and clean handoffs to enterprise systems and human care.
Saathi Assist is a support and guidance companion. It is not a replacement for doctors, therapists, emergency medical care, or clinical diagnosis.
Built for governed deployment.
The hygiene that makes a care platform safe to roll out across a workforce.
Care boundary enforcement
Saathi is built as a support companion. Boundaries against diagnosis or prescription are enforced in the conversation layer, not left to prompt discipline.
Row-level security on user data
Each user's records are isolated at the database layer. Access is scoped per-user by default.
Encrypted credentials & secrets
API keys, model credentials, and integration secrets are stored encrypted, never in client code.
Session & audit logging
Conversations, escalations, and document approvals are logged for clinical and governance review.