-
-
Main AI companion chat with voice, persona, breathing, and quick actions.
-
Companion picker showing specialized AI personas.
-
Daily Pulse check-in for mood, energy, and user needs.
-
Smart continuity showing saved threads and next steps.
-
Moment actions turning AI replies into plans, proof cards, and capsules.
-
Parent-only fast access tools.
-
Homework transformer upload flow.
-
Child profile separation under one parent account.
-
Gemini homework transformation result.
-
Child workspace home with “Hi Rana.”
-
Interactive child homework activity.
-
Story journey passport and progress.
-
Interactive story reader with quiz and guide options.
-
Child-safe companion picker.
-
Parent pulse snapshot, return code, and oversight.
Inspiration
FadFada | فضفضة was inspired by a simple human need: everyone needs a safe space to talk, feel understood, and take one small next step.
In Arabic, “فضفضة” means opening up, venting, and letting what is inside come out. Many people carry stress, loneliness, burnout, family pressure, study pressure, or private thoughts they cannot easily share. Parents face another challenge: they want to support their children with learning and emotions, but they need AI tools that are safe, age-aware, and clearly separated from adult accounts.
FadFada was created as a calm Arabic-first bilingual AI companion for both personal reflection and family support. It gives adults a private space to reflect, while also giving children a safe, separate workspace for stories, homework, learning, drawing, music, and guided conversation.
The goal is not to replace therapy, teachers, or parents. The goal is to create a trusted first space: a place to express, organize, learn, and move forward safely.
What it does
FadFada is an Arabic-first bilingual AI reflection and family companion powered by Gemini.
Users can chat in Arabic or English, choose emotional modes and companion personas, receive supportive reflection, save important moments, create proof cards, start small personal challenges, and turn meaningful conversations into useful artifacts.
For families, FadFada adds role-aware parent and child workspaces. Parents can create child profiles, set limits, upload homework images, and transform worksheets into child-ready activities. Children get their own safe space with child-friendly companions, stories, quizzes, voice answers, interactive homework games, drawing, daily moments, and a color piano.
FadFada includes:
- Arabic and English support
- RTL/LTR interface
- Gemini-powered reflection
- multiple companion personas
- emotional worlds and modes
- saved moments and memory snapshots
- proof cards and tiny plans
- growth challenges
- parent, child, and admin role experiences
- child-safe companion roster
- parent homework transformer
- Gemini multimodal worksheet understanding
- structured JSON activity generation
- child-safe stories, quizzes, and guided learning
- child drawing board
- child color piano / music activity
- child daily moments
- parent oversight, activity signals, and return-code control
- admin evidence and operations dashboard
- companion quality ratings
- PWA support for mobile and desktop
FadFada is not therapy, medical care, diagnosis, or emergency support. It is a safe first space for emotional reflection, family learning support, and self-organization.
How we built it
We built FadFada as a responsive Progressive Web App using:
- Next.js
- React
- TypeScript
- Tailwind CSS
- Prisma
- Neon Postgres
- NextAuth
- Gemini / Vertex AI
- Google Cloud Text-to-Speech integration path
- Vercel
- PWA manifest and service worker
- bilingual Arabic/English UI
- RTL/LTR layout support
The app includes production-style features for user accounts, parent and child workspaces, admin monitoring, chat sessions, profile data, AI reflections, homework transformation, child-safe prompts, safety-aware responses, and evidence collection.
Gemini is used in several ways:
- text reasoning for companion replies
- persona-aware and mode-aware responses
- multimodal understanding for homework images
- structured JSON outputs for child activities
- role-aware prompts for parent and child separation
- child-safe guided learning flows
- parent summaries and activity transformation
A key part of the architecture is workspace identity. Parent and child experiences are separated server-side, not only visually. The app derives the active role from the authenticated session and workspace context, so the AI knows whether it is speaking to a parent or a child.
Challenges we ran into
The biggest challenge was balancing emotional usefulness with safety.
FadFada deals with sensitive feelings, so we avoided positioning it as an AI therapist or medical tool. We focused on supportive reflection, careful wording, and guidance toward professional or emergency help when needed.
Another challenge was building a product that works well in both Arabic and English. This required more than translation. Direction, tone, layout, cultural meaning, typography, and mobile usability all mattered.
The most important technical challenge was parent and child context isolation. A family account can contain both parent and child spaces, but the AI must always know who it is speaking to. We solved this with server-side workspace identity, separate child-safe prompts, scoped child context, and leak-guard observability so parent identity does not enter the child experience.
We also had to make the app feel complete across three roles: parent, child, and admin. Each role needed a different interface, different permissions, and different AI behavior.
Accomplishments that we're proud of
We are proud that FadFada became more than a generic chatbot.
It is Arabic-first, emotionally aware, culturally meaningful, mobile-friendly, and powered by Gemini. We built real product workflows: reflection, saved moments, proof cards, challenges, account profiles, parent tools, child profiles, homework transformation, child-safe companions, child stories, drawing, music, admin analytics, and an evidence dashboard.
We are especially proud of the parent and child experience. A parent can upload a worksheet, Gemini can understand it, and FadFada can turn it into a child-ready activity with choices, hints, and guided steps.
When the child enters their workspace, the app changes identity completely: child-safe companions appear, adult tools disappear, and the AI speaks to the child using the correct name, age, and safety context.
We are also proud of the return-code control. A child can enjoy a protected space, while returning to parent tools requires parent confirmation.
Finally, we are proud of the name. “FadFada” carries emotional meaning in Arabic while still being memorable globally.
What we learned
We learned that emotional AI products need trust before features.
Key lessons:
- Arabic-first UX needs cultural care, not only translation.
- Safety boundaries are essential.
- Parent and child AI spaces must be isolated by design.
- Gemini is strongest when it powers real workflows, not only chat.
- Multimodal AI can turn static homework into interactive learning.
- Structured JSON makes AI outputs reliable inside a product.
- Admin evidence matters for real product validation.
- Mobile polish is critical for users and judges.
- A family AI product needs different experiences for parent, child, and admin roles.
- Trust depends on clear boundaries, not only good responses.
Revenue and business model
Current revenue is $0.
FadFada is still at the hackathon and early product validation stage. Our focus so far has been building a working product, validating the core family workflows, and preparing for real user feedback.
The planned business model is a freemium family subscription:
- free tier for basic reflection and limited child profile access
- paid family plan for multiple child profiles, homework transformation, saved history, reports, and advanced tools
- future school or organization pilots for supervised learning and family support workflows
- optional premium voice, media, and advanced AI features
Our near-term revenue goal is to validate pricing through pilot families, waitlist demand, and paid beta interest.
What's next for FadFada
Next, we want to grow FadFada into a real multilingual emotional reflection and family learning platform.
Planned next steps:
- improve voice and video reflection
- add official cloud speech-to-text and text-to-speech
- expand safety classification
- improve long-term memory controls
- strengthen parent reports and child progress insights
- improve homework follow-up workflows
- add more child-safe learning activities
- collect real user feedback
- launch a paid beta
- validate family subscription pricing
- support more languages and cultures
- explore school and family-support partnerships
The long-term vision is for FadFada to become a trusted AI companion that helps people express themselves, supports families safely, and helps children learn in a protected, age-aware space.
Built With
- ai
- css
- flash
- gemini
- generative
- json
- neon
- next.js
- nextauth
- postgresql
- prisma
- pwa
- react
- recognition
- rtl
- speech
- tailwind
- text-to-speech
- typescript
- vercel
- vertex

Log in or sign up for Devpost to join the conversation.