🧠 REFRAME

AI-powered communication safety for healthier digital conversations

REFRAME is a functional, startup-grade prototype built for MINDCODE 2026 (HealthTech Hackathon).
It helps people rewrite emotionally risky messages before sending, reducing escalation, stress, and misunderstanding across workplace and online messaging platforms.

---

🌍 The Communication Problem

Work and online communities now run on fast text interactions:

• Slack in deadline crunches
• Discord in high-emotion discussion channels
• Customer support tools under response-time pressure

In these moments, a message that is technically correct can still feel harsh, dismissive, or hostile.
Tone misunderstanding creates emotional strain and professional risk:

• 😟 anxiety from overthinking every message
• ⚡ conflict from accidental escalation
• 🧠 cognitive load from constant emotional self-editing
• 🔥 burnout from repeated communication friction

REFRAME’s Core Intervention

REFRAME helps before send-time, not after damage. It suggests healthier wording that:

• preserves intent
• reduces emotional intensity
• improves clarity and social safety

Small daily improvements in language compound into healthier team dynamics and lower stress over time.

---

🧠 Why This Matters

Real-world pressure scenarios

• Customer support pressure: agents responding quickly may write in ways perceived as cold or dismissive.
• Deadline conflicts: urgency can produce abrupt language that escalates team friction.
• Tone ambiguity: text lacks vocal cues, so intent is often misinterpreted.

These micro-escalations compound into larger outcomes: team distrust, emotional fatigue, and professional risk.

Preventive Wellbeing Layer

REFRAME does not replace therapy, human empathy, or judgment.
It adds a practical, always-available layer of communication support where stress actually happens: everyday messaging interfaces.

---

⚙️ How It Works

End-to-end flow

1. ✍️ User drafts a message in a demo communication interface.
2. 🔐 Input is sanitized and processed through secure server-side API routes.
3. 🧭 AI detects emotional escalation risk (low, medium, high).
4. 🪄 REFRAME generates tone-safe alternatives.
5. 📈 Stress Reduction Meter reflects potential emotional impact improvement.
6. 🔎 Transparency features explain why suggestions are shown.
7. ✅ User chooses whether to keep original or send a reframed version.

Explainability-first behavior

• Suggestions are assistive, not forced.
• Users remain in full control.
• UI communicates both confidence and limitations.

---

✨ Product Features

Communication Intelligence

• 🪄 Inline tone rewriting directly in chat flow
• 🚨 Real-time escalation detection for risky language
• 🧠 Context-aware reframing that preserves original intent

Multi-Environment Demo Hub

• 💼 Slack-style workspace simulation
• 🛰️ Discord-style community simulation
• 🎧 Customer support portal simulation
• 📱 Mobile Figma prototype entrypoint

Responsible AI Safeguards

• 🔒 Server-only AI execution
• 🧱 Prompt injection defenses
• 🧼 Input sanitization and controlled formatting
• ⛔ Anonymous usage caps and abuse throttling
• 📜 Transparent non-clinical AI framing

AI Provider Strategy

• ✅ OpenAI integration active in current runtime
• 🧭 Gemini fallback planned as roadmap failover

---

💚 Mental Health Impact

REFRAME is designed around preventative wellbeing, not treatment.

Potential impact areas:

• Reduced psychological stress: less emotional reactivity in high-pressure chat moments.
• Improved communication confidence: users learn safer phrasing patterns over time.
• Healthier workplace dynamics: fewer misunderstandings and lower escalation frequency.
• Lower cognitive load: AI-assisted phrasing reduces overthinking before sending.
• Escalation prevention: supports calmer communication loops early, before conflict deepens.

By improving everyday language habits, REFRAME aims to help teams and communities communicate with more emotional safety and resilience.

---

🎮 Demo Experience

Explore REFRAME in realistic communication contexts:

• Web demos: interactive Slack-style, Discord-style, and support-style environments
• Mobile demo: opens an external Figma prototype experience

🔗 Live Demo Links

• 🌐 Website Demo: REFRAME Web App
• 📱 Mobile Prototype (Figma): Open Figma Mobile Demo

What reviewers can test

• Emotional risk classification output
• Suggested rewrites with intent preservation
• Stress meter response to wording changes
• Transparency panel explanations

---

🔒 Security & Ethical Design

REFRAME follows privacy-first health-technology principles:

• 🔐 API keys remain server-side only
• 🚫 No client-side key leakage
• 🧱 Prompt injection resistance via sanitization and controlled prompts
• 🚦 Rate limiting and cooldown safeguards
• 👤 Anonymous usage restriction without mandatory sign-in
• 🧾 Transparent AI boundaries (assistive, non-diagnostic)

---

🚀 Getting Started

1) Clone

git clone https://github.com/DhruvP2205/Reframe.git
cd Reframe

2) Install dependencies

npm install

3) Configure environment variables

cp .env.example .env.local

.env.local:

OPENAI_API_KEY=your_openai_api_key
GEMINI_API_KEY=your_gemini_api_key_optional

.env.example currently includes OpenAI by default. Gemini is optional and roadmap-oriented.

4) Start development server

npm run dev

Open: http://localhost:3000

5) Validate

npm run lint
npm run build

---

🗂️ Project Structure

reframe/
├─ src/
│  ├─ app/
│  │  ├─ api/analyze/         # Secure AI analysis endpoint
│  │  ├─ demo/                # Demo hub + simulation pages
│  │  ├─ about/               # Team / mission page
│  │  ├─ settings/            # User preferences + controls
│  │  ├─ layout.tsx
│  │  └─ page.tsx
│  ├─ components/             # UI components (chat, carousel, meters)
│  ├─ data/                   # Demo conversation seed data
│  ├─ lib/                    # Shared helpers, prompts, env utilities
│  ├─ server/                 # Security, sanitization, rate limit logic
│  └─ types/                  # Shared type contracts
├─ public/
├─ LICENSE
└─ README.md

---

📊 Usage Limits

To keep prototype behavior stable and fair:

• Rate limit per IP
• Cooldown between rapid requests
• Duplicate-message rejection
• Anonymous usage caps per visitor/session

These controls support both safety and infrastructure reliability.

---

⚠️ Known Limitations

• Prototype-first scope; not yet enterprise production software
• AI output can be imperfect and requires human review
• No emergency mental health escalation protocol
• Gemini fallback is not active in current runtime (roadmap only)

---

🔭 Future Vision

• 🔁 Full OpenAI ↔ Gemini automatic failover
• 🎯 Personal tone profiles by team role/context
• 📈 Communication health analytics dashboards
• 🌐 Multilingual reframing support
• 🧩 Native integrations for Slack/Discord/helpdesk platforms

---

🤝 Contributing

Contributions are welcome and appreciated.

1. Fork the repository
2. Create a feature branch (feat/your-feature-name)
3. Commit with clear, focused messages
4. Run npm run lint and npm run build
5. Open a PR with context, rationale, and screenshots for UI changes

Please align contributions with REFRAME’s privacy-first, ethical-AI mission.

---

📄 License

This project is licensed under the MIT License.

See LICENSE for full terms.

---

👨‍💻 Developer Credits

Built by beTheNoob for MINDCODE 2026.

• ▶️ YouTube: https://www.youtube.com/@beTheNOOB
• 🐙 GitHub: https://github.com/DhruvP2205
• 💼 LinkedIn: https://www.linkedin.com/in/dhruv-prajapati-cybersecurity/

---

🫶 REFRAME Mission

Reduce communication-driven mental stress through thoughtful, transparent, and responsible AI assistance.