# Peekaboo

- **Event:** [fal x Sequoia 72-Hour Video Hackathon](https://cerebralvalley.ai/e/fal-x-sequoia-virtual-hackathon)
- **When:** Jul 17 at 12:00 PM – Jul 19 at 8:00 PM (PDT)
- **Where:** Online
- **Team:** [Rahil Singhi](https://cerebralvalley.ai/u/Rahilsinghi)
- **GitHub:** https://github.com/rahilsinghi/peekaboo
- **Demo video:** https://youtu.be/KhVyKZgbPTQ
- **Gallery:** https://cerebralvalley.ai/e/fal-x-sequoia-virtual-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/fal-x-sequoia-virtual-hackathon/hackathon/gallery/89

Peekaboo is a live, perceptive cartoon tutor for preschoolers. Kids' shows pretend to listen — Dora asks, waits, and moves on regardless. Peekaboo actually sees the child: MediaPipe perception runs locally at 70fps reading fingers, nods, head-shakes, and attention, and a pure state-machine episode engine only advances when the child truly answers — hinting gently after misses, and noticing when they look away (Boo hides and calls them back). Every episode is a generated claymation film: an original owl character locked with a FLUX.2 LoRA performs a 45-clip library (Kling o3, Veo 3.1), speaks with one ElevenLabs voice, and teaches counting inside themed worlds — pick space, dinosaurs, ocean, or candy, or type anything and watch clay sprites generate live on fal's queue mid-episode. The child even joins the cast as a generated clay cameo. It solves passive screen time: this is a screen that genuinely responds, teaches, and adapts. Built solo, from scratch, in 44 hours, on fal.

---

Markdown version of https://cerebralvalley.ai/e/fal-x-sequoia-virtual-hackathon/hackathon/gallery/89. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt