HackStack

Project Info

Dreamscape

Devpost

Inspiration

(ง •̀_•́)ง Dreams are often left forgotten, fading with the morning light before they can be understood or remembered. Most vivid dreams happen during REM sleep, a stage of the sleep cycle that typically occurs every 90 minutes - when your brain is highly active, similar to when your awake. Because the brain's memory-consolidation systems are not fully active during REM, dreams are lost within minutes of waking. That’s why we built Dreamscape, a web app that captures your dreams through voice recording and brings them to life with AI-generated video, preserving the emotions and memories before they're gone

What it does

(｡•̀ᴗ-)✧ When users wake, they can open Dreamscape to instantly record a voice memo capturing the details of their dream. Once the recording is complete, an AI-generated video brings the experience to life. Each entry is saved in a personal gallery, allowing users to revisit and reflect on the stories their minds create while they sleep.

How we built it

( •̀ .̫ •́ )✧ Set up a STT model, set it up in parallel and ran 5-second chunk inference to detect emoji, and total STT inference to obtain raw transcript Translate and detail the raw transcript into JSON and prompt engineer to Veo 2 Veo 2 takes in this output and outputs video based on preferences (optional) + prompt

Challenges we ran into

(ノ°Д°)ノ We didn't have access to Veo 3 - most of our prompt-to-video generation was fairly behind Google API keys had to be rotated constantly due to usage limit

Accomplishments we're proud of

(๑˃̵ᴗ˂̵)و The UI of the Dreamscape looks very ethereal 😻 - it captures the essence of a dream which is what we initially wanted to go for. We bootstrapped fairly fast and led UI overhaul for the last 6 hours of the hackathon.

What we learned

(•‿•) We both learned how to connect Veo 3 with NexTJS apps. Jason I learned how to cheat Google by using all 10 of your links for credits. Additionally, I learned how to run tasks in parallel and persist components across states. Christine Got more experience with Next.JS and implementing the front-end :D

What's next

(¬‿¬) For future iterations, Dreamscape would be adapted into a mobile app for easier access for users to easily record their dreams. More specifically, when users would prompt alarms/use action buttons on iOS devices, we would want Dreamscape to take in all notes. For an unrealistic expectation, we would want to connect to users physically through hardware throughout the night and interpret dreams.

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • Next.jsIn code
  • ReactIn code
  • SupabaseIn code
  • Tailwind CSSIn code
  • TypeScriptIn code
  • OpenAIClaimed

6 of 7 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

86 KB

Source files

23

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars