Hypnos

Hypnos lets you speak your dreams before they fade. An AI voice turns them into clear takeaways and quietly stores them, giving your subconscious a place to stay.

Comment

“Good morning,

Sleep well? What was the dream?

Oh. Gone already? Yeah, that checks out. I’m Hypnos. I help you remember what your brain clearly did not plan to keep.

Talk while it’s still fuzzy.

Tell me whatever you recall — even the weird parts that make no sense. I’ll listen, pull out what actually matters, and reflect it back in a way that’s easy to understand. Nothing mystical. Just a clearer look at what your mind came up with.

Each dream is distilled down and set aside, quietly marked by time, should it ever be needed again.“

🌙 The Story of Hypnos

🚆 What Inspired Us

Hypnos began on a train ride to Kingston.

One of our team members fell asleep somewhere between stations and dreamed they were drinking water alongside dinosaurs — calm, surreal, and completely confusing. When they woke up, the question was immediate:

“What on earth did that mean?”

Like most dreams, it vanished before it could be understood. That moment — watching a strange, vivid experience dissolve in seconds — sparked the core idea behind Hypnos:
What if your dreams didn’t have to disappear?

We realized that dreams sit in a strange space between memory, emotion, and meaning. They feel important, but we lack a simple way to capture them before the brain overwrites them. Hypnos was born as a way to preserve those fleeting moments — not to diagnose or analyze, but to reflect them back to you while they’re still alive.

🧠 What We Learned

We learned two big things while building Hypnos:

Dreams fade almost immediately after waking, so capturing them has to be fast and effortless. Voice input turned out to be the most natural solution.
People don’t want their dreams analyzed or diagnosed — they want them reflected back in a calm, thoughtful way that helps them make sense of what they experienced.

How We Built It

🛠 How We Built It

Hypnos is built as a voice-first mobile experience:

  • Text-to-speech with the ElevenLabs captures the dream the moment the user starts talking
  • An AI summarization and theme extraction layer and speech-to-text with Gemini API distills the dream into its key ideas, reflecting it back in a mysterious, wise voice.
  • A dream journal database stores each summarized dream, organized by date

The pipeline looks roughly like: [ \text{Voice} \rightarrow \text{Text} \rightarrow \text{Themes} \rightarrow \text{Narration} \rightarrow \text{Journal} ]

Everything is designed to minimize friction between waking up and capturing the dream.

⚔️ Challenges We Faced

The hardest part wasn’t the tech — it was the experience.

  • How do you make AI feel calm instead of clinical?
  • How do you make the interpretation feel thoughtful?
  • How do you summarize something as abstract as a dream without flattening it?

We also had to deal with messy speech input, fragmented memories, and people jumping between ideas mid-sentence. Dreams don’t arrive in clean paragraphs — they arrive like static. Our system had to learn to find signal inside that noise.

🌌 Why Hypnos Matters

It all started with dinosaurs on a train.

That one forgotten dream reminded us how much of our inner world disappears every morning. Hypnos exists to hold onto those moments — not to explain you, but to let you hear yourself.

Sometimes that’s all a dream really needs.

Built With

Share this project:

Updates