Sentient AI for Knowledge, Support & Healing Interactions
Your Personal AI Therapist — Empathetic, Confidential, Always Available.
SAKSHI AI is an AI-powered therapeutic companion designed to understand human emotions, simulate real therapist interactions, and support users in navigating their mental well-being. Built with empathy at its core, SAKSHI listens, learns, and responds like a therapist — providing a safe, supportive environment for self-reflection and emotional growth.
-
🗣️ Natural Conversations
Engage in human-like dialogue with an intelligent, responsive AI therapist. -
🧠 Emotion Recognition (In Progress)
Understands emotional cues from your language and responds accordingly. -
💬 Reflective Feedback Generation (Planned)
Offers insights and therapeutic responses that help users reflect and grow. -
🔐 Privacy-First
All interactions remain local. No data is stored on the cloud. -
📖 Session Continuity
Maintains conversation context for seamless therapy-like experience.
- Individuals seeking low-pressure, AI-assisted mental health support.
- People interested in emotional journaling or self-reflection.
- Users looking for a confidential, always-on companion.
- 🌙 End-of-day emotional check-ins
- 💡 Self-awareness training
- 🎯 Goal setting & tracking mental clarity
- 😔 Coping support during difficult times
Include user interface screenshots or diagrams here to show what the app looks like.
This section is for contributors or developers looking to understand the architecture, flow, and how to run or extend SAKSHI AI.
| Layer | Technology |
|---|---|
| Frontend | AngularJS |
| Backend | Python (Flask - if used, TBD) |
| Execution | CLI (main.py) |
| Hosting | Localhost / On-Prem |
sakshi-ai/
│
├── backend/
│ ├── main.py # App entry point
│ ├── session_manager.py # Session tracking per user
│ ├── emotion_detector.py # (Planned) Emotional NLP processing
│ ├── feedback_generator.py # (Planned) Therapist-style response engine
│ └── utils/ # Utilities and helper methods
│
├── frontend/
│ └── [AngularJS App Files] # Web-based UI
│
├── assets/
│ └── logo/, samples/, etc.
│
├── README.md
└── config/
└── settings.json # (Optional) Runtime configuration
- User Input is entered through the AngularJS frontend UI.
- Backend API (
main.py) receives the text input. - Session Manager checks if it's a new or existing session.
- (Future) Emotion Detector analyzes emotional tone.
- (Future) Feedback Generator builds a reflective, helpful response.
- Response is returned to frontend and displayed to user.
- Conversation state is stored locally to preserve context.
cd backend/
python main.pyEnsure Python 3.9+ and required dependencies are installed.
Serve via any static web server or lite-server:
cd frontend/
lite-server- Emotion analysis using sentiment NLP
- Reflective feedback with therapy scripts
- Voice-to-text & text-to-voice integration
- Local secure chat history
- Therapist session timeline visualization
We welcome developers, therapists, and researchers!
- Fork the repo
- Create a feature branch
- Submit a well-documented PR
- All conversations are stored locally, not transmitted to third parties.
- Future versions will comply with HIPAA-like standards for data privacy.
- No behavioral data is used for training external models.
License type to be finalized. All rights reserved until open-sourced.
SAKSHI AI is developed by a team passionate about mental health, AI, and empathetic design.
Created & Maintained by Sraban Kumar Patro & Saroj Kumar Sahoo Contact: reg.sraban@gmail.com
- 📚 External Libs/Models: TBA
- 🧠 Emotion Dataset Source: TBA
- 🔍 Logs / Debugging:
logs/or browser console - 🔒 Secure Config: Move secrets/envs to
.env(Planned)