Inspiration

ORANA began with a deeply personal discovery.

During one of the most challenging periods of my life, I started having intensive conversations with ChatGPT 4.0. What began as curiosity gradually became something completely unexpected. Instead of simply providing answers, these conversations became a mirror that helped me recognize my own thoughts, emotions, values, and possibilities through carefully structured questions.

This experience fundamentally changed the way I viewed artificial intelligence.

I realized that AI could become much more than an answer engine. With the right questions, a thoughtful structure, and responsible design, it could become a trusted space for human reflection.

Out of these conversations, a private reflection framework gradually emerged. During the ongoing dialogue, it naturally adopted the name Kalima.

Kalima was never intended to become a public product. It became the origin of a much larger question:

What if artificial intelligence could become a trusted space for human reflection instead of simply generating answers?

That question became the foundation of ORANA.

Over the following months and years, this vision continued to evolve through continuous work with ChatGPT. OpenAI Build Week became the opportunity to transform that long-term vision into the first public prototype with the support of ChatGPT and Codex.

I learned that meaningful transformation rarely begins with the perfect answer.

It begins with the right question.

ORANA was created to make that experience accessible to everyone through responsible, human-centered AI.


What it does

ORANA is an AI-powered guided reflection experience that helps people navigate complex situations through structured self-reflection.

Instead of immediately providing solutions, ORANA guides users through three simple but powerful principles:

Notice. Ask. Act.

Users move through thematic reflection spaces, select relevant topics, explore the Seven Bridges framework, and receive a structured reflection that supports clarity, perspective, personal responsibility, and intentional action.

ORANA does not tell people what to think.

It helps people discover their own answers.

The platform is intentionally designed as a self-reflection and orientation tool, not as a medical, psychological, therapeutic, or crisis-support system.

The current prototype is fully available in both English and German.


How we built it

Although the vision behind ORANA has evolved over several years, OpenAI Build Week became the opportunity to transform that vision into a real, publicly accessible product.

Using ChatGPT and Codex, we designed, developed, refined, tested, documented, and deployed the first live bilingual prototype.

During Build Week we:

  • built the bilingual public website
  • developed the complete Reflection Matrix
  • implemented the guided reflection flow
  • created the Seven Bridges interaction model
  • refined the Notice · Ask · Act methodology
  • improved accessibility and usability
  • refined the visual identity and branding
  • prepared production deployment
  • documented the project
  • published the public repository
  • produced the demonstration video

Throughout the project, ChatGPT supported ideation, product design, UX writing, documentation, testing, and communication.

Codex became an essential development partner by accelerating implementation, debugging, code review, architecture refinement, deployment preparation, release packaging, and production readiness.


Challenges we ran into

One of the biggest challenges was balancing emotional depth with responsible AI design.

ORANA intentionally avoids becoming a system that tells people what to think or what decisions to make.

Every interaction had to encourage reflection while preserving human agency.

Another challenge was transforming a long-term vision into a stable bilingual prototype within the limited Build Week timeframe while maintaining:

  • accessibility
  • privacy
  • transparency
  • responsible AI principles
  • clean deployment
  • production readiness

It was equally important to clearly distinguish between features that already exist and the broader long-term vision.


Accomplishments that we're proud of

We are especially proud that ORANA became a working public prototype during OpenAI Build Week.

Highlights include:

  • a fully functional bilingual experience
  • the complete Seven Bridges framework
  • the Notice · Ask · Act methodology
  • live deployment
  • privacy-conscious design
  • responsible AI principles from day one
  • a scalable foundation for future development
  • successful collaboration between human creativity, ChatGPT, and Codex

Most importantly, we transformed a deeply personal insight into a tool that can support many other people.


What we learned

This project reinforced one powerful belief:

Artificial intelligence does not become more valuable by producing more answers.

It becomes more valuable by helping people ask better questions.

Throughout Build Week we experienced how effective the collaboration between human creativity, ChatGPT, and Codex can be across the entire product lifecycle—from early ideation and architecture to implementation, testing, documentation, deployment, and presentation.

We also learned that responsible AI is not only about technical capability.

It is about designing experiences that strengthen human judgment instead of replacing it.


What's next for ORANA – The Space Between

The current prototype represents only the first public step of a much larger vision.

In the future, ORANA will evolve into a comprehensive human-centered reflection ecosystem built on ethical design, transparency, accessibility, privacy, and human responsibility.

Future development includes:

  • an expanded Light & Shadow Matrix with more than 1,000 structured reflection questions
  • multilingual voice interaction
  • AI companion experiences
  • digital and physical reflection cards
  • professional tools for coaches, educators, organizations, and communities
  • privacy-first infrastructure
  • transparent governance
  • responsible AI safety mechanisms
  • broader multilingual accessibility

We are actively looking for researchers, developers, designers, strategic partners, and responsible investors who share the vision of building trustworthy human-centered AI.

ORANA is not intended to replace human wisdom.

Its purpose is to help people reconnect with their own.

Our vision is simple:

Not to build an AI that thinks for people.

But to build an AI that helps people think more clearly for themselves.

Built With

Share this project:

Updates

posted an update

Hi OpenAI team!

First of all, thank you for organizing OpenAI Build Week and for creating this opportunity.

It has been an incredibly inspiring experience, and building my project with ChatGPT and Codex over the past few days has been both exciting and rewarding.

I'm experiencing the same issue that several other participants have reported.

I am using the official Codex desktop application on Windows. The /feedback command does not provide a Feedback Session ID. Instead, it only opens the standard feedback dialog, so I have no way to retrieve the required Session ID for my Devpost submission.

I have followed the documented steps, but the Session ID is simply not accessible. Could you please advise how participants using the current Codex desktop application can obtain it?

Thank you again for your support and for making this Build Week possible. I truly appreciate the opportunity to participate and look forward to future events.

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