NarraTrail

NarraTrail turns any place into an AI-powered educational adventure, transforming local heritage and family exploration into immersive, playable stories.

Comment

Inspiration

NarraTrail began in Zumel, a small village in Burgos, Spain, while I was creating a real-world adventure for children.

I photographed the village streets, church, old cemetery entrance, stone features, and other details that normally go unnoticed. I then transformed those real locations into clues, challenges, and pieces of a larger story.

That first experience became a playable offline gymkhana:

Try the original Zumel adventure prototype

Testing it with children showed me that a village does not need a theme park to become exciting. It already contains characters, mysteries, history, landscapes, and unexplored paths. The children were not simply looking at the village anymore: they were investigating it.

However, creating the experience manually required local research, photography, writing, image preparation, programming, narration, and repeated testing. That led to the central idea behind NarraTrail:

What if families, teachers, and local communities could turn any real place into a personalised educational adventure?

What NarraTrail does

NarraTrail transforms real locations into interactive, story-driven educational trails.

Parents, teachers, and local creators can select what they want children to develop—such as mathematics, history, science, reading comprehension, spatial orientation, social skills, or teamwork—and create an adventure adapted to their age, interests, and surroundings.

Players discover locations progressively, solve physical and digital challenges, listen to narration, and watch familiar places become part of a fictional world. Undiscovered locations remain hidden, preserving mystery without revealing future stages.

The first NarraTrail demonstration, The Legacy of Sancha, is based on real locations in Zumel, including its square, church, old cemetery entrance, stone features, and surrounding landscape.

Historical facts, local traditions, legends, and fictional elements are clearly separated so that imagination enriches the place without misrepresenting its history.

How I built it

I used OpenAI Codex as a development partner throughout an iterative, real-world design process.

Rather than beginning with an abstract specification, I repeatedly tested the experience on a mobile phone while walking through the actual locations. I recorded voice notes, photographed environmental details, identified possible clues, and revised the interface and story according to what worked in the real world—not just behind a desk.

Codex helped me transform those observations into structured requirements, organise the project, improve the interface, implement new functionality, identify technical problems, and test successive versions.

The project combines:

  • A mobile-first player experience.
  • English and Spanish language support.
  • A creator studio for building new adventures.
  • Educational objective, age, and difficulty settings.
  • Individual configuration and media for each location.
  • Maps and orthophotos connected to story progression.
  • Hidden, locked, and discovered location states.
  • A Game Master mode for supervising the experience.
  • Safety and accessibility information.
  • Importable and downloadable adventure packages.
  • An offline-capable demonstration that does not require an API key while playing.

For narration, I designed a generation workflow using the OpenAI Speech API. The objective is not to produce a theatrical or artificial narrator, but a warm and natural voice that sounds as though someone were personally telling a story to a child.

The audio is generated in advance and included in the adventure package. This keeps API credentials outside the browser and allows the experience to remain playable offline.

The source code, development history, documentation, and Build Week materials are available here:

Explore the NarraTrail source code and documentation

Challenges I faced

The greatest challenge was connecting a digital narrative to an unpredictable physical environment.

A conventional game controls its entire world. NarraTrail must work with real distances, weather, accessibility, private property, road safety, monuments, and places where children must not climb, enter, or move objects.

Safety therefore had to become part of the product architecture rather than a final warning screen.

Another challenge was balancing simplicity and creative control. A creator needs enough options to build a meaningful adventure, but parents and teachers should not face an intimidating professional editor. This led to a progressive interface with accessible default options and more detailed controls when needed.

I also had to preserve mystery without confusing players. Early versions revealed too much information on the map, so the progression system was redesigned to display only locations that players had legitimately discovered.

Another important challenge was maintaining consistency between real photographs, generated images, maps, narration, historical information, and fictional elements.

Finally, producing natural narration proved more difficult than expected. Local text-to-speech tools sounded robotic and damaged the emotional experience. Using prompted speech generation through the OpenAI API made accent, pace, warmth, intonation, and pronunciation part of the creative process.

What I learned

The most important lesson was that AI becomes much more valuable when it is grounded in a real place and guided by human knowledge.

AI can help structure stories, adapt difficulty, generate alternatives, organise assets, and accelerate development. However, the creator still provides the local context, educational intent, safety judgement, historical responsibility, and emotional meaning.

I also learned that prototyping in the actual environment changes the product. Features that appear clear on a desktop may become confusing while walking, holding a phone, supervising children, or trying to locate a physical clue.

Most importantly, I learned that rural locations should not be treated as empty spaces waiting for technology. Their architecture, memories, landscapes, and small details are already rich interfaces. NarraTrail simply helps people notice them.

What's next

The next step is to evolve NarraTrail from a single-location demonstration into a platform where families, schools, cultural organisations, and rural communities can create, exchange, and play their own adventures.

Future development will focus on:

  • Faster AI-assisted adventure creation.
  • More educational objectives and age adaptations.
  • Improved accessibility alternatives.
  • Reusable community-created adventure packages.
  • Collaborative and multiplayer experiences.
  • Tools for schools, municipalities, museums, and rural tourism organisations.
  • Support for connected routes between neighbouring villages.

The long-term goal is simple:

NarraTrail. Turn any place into an adventure.

Built With

Share this project:

Updates