SaneSpaces

Inspiration

We were inspired by a simple question: Why should emotional support require people to abandon their language, culture, or natural way of speaking? While today's AI assistants can communicate in many languages, they often treat language as translation rather than part of the care experience. We set out to build a wellness companion that feels familiar, culturally authentic, and safe.

What it does

SaneSpace is a multilingual, culturally adaptive AI companion that offers users a private space to reflect, seek guidance, and have supportive conversations. It adapts to users' language, communication style, and emotional context while using a deterministic safety pipeline to identify high-risk situations and provide appropriate responses.

How we built it

We built SaneSpace using Next.js 14, TypeScript, Tailwind CSS, and Framer Motion. AI conversations are powered through server-side API routes, while our orchestration layer determines language, communication style, voice provider, and safety response before delivering each interaction. Nigerian voices are handled by YarnGPT, international voices by ElevenLabs, and the platform includes automatic fallback to device speech if premium voice services become unavailable.

Challenges we ran into

Our biggest challenge was building an experience that felt culturally authentic while remaining scalable. We also had to design an orchestration layer that could intelligently route between multiple voice providers, support different communication styles, and prioritize user safety—all within the limited timeframe of a hackathon.

Accomplishments that we're proud of

Built a working multilingual AI wellness companion in just a few hours. Integrated multiple voice providers behind a single intelligent routing system. Delivered authentic Nigerian voice experiences instead of generic multilingual voices. Implemented Care Mode with region-specific crisis resources. Created a personalized experience through language, voice, communication style, and emotional continuity.

What we learned

We learned that building a great AI wellness experience is about much more than choosing a powerful language model. Cultural authenticity, voice, personalization, safety, and thoughtful orchestration all play a critical role in creating conversations that feel supportive and trustworthy.

What's next for SaneSpaces

Our next steps are to expand language and voice support, strengthen emotional continuity across conversations, introduce secure cloud-based user profiles, launch mobile applications, and partner with organizations to provide more localized wellness and support resources. Our architecture is designed to scale globally by adding new languages, voice providers, and regional safety resources without rebuilding the core conversation system.