# Team Ayush

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Ayush Khanal](https://cerebralvalley.ai/u/ayush2222)
- **GitHub:** https://github.com/ayushkhanal1/Echoed
- **Demo video:** https://youtu.be/0jkxiCp4QYY
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1941

Echoed is a privacy-first platform that connects strangers who independently experience identical, hyper-specific coincidences—such as thinking of an old friend right before an unexpected call, waking at the exact same odd hour to identical sensations, or experiencing parallel deja vu, without either person ever searching for the other. Users document personal experiences into private vaults. Echoed pairs a native PostgreSQL trigram recall engine (<5ms) with Meta Muse Spark Open (`meta-models/muse-spark-open-1026` on Baseten) to analyze semantic, contextual, and temporal convergence, score coincidence confidence, filter PII, and generate synthesized thread summaries.

The Problem It Solves:
1. Isolation & Ridicule: Profound, uncanny coincidences happen every day, but people keep them to themselves out of fear of skepticism, stigma, or disbelief.
2. Fabrication on Social Media: Public forums and social feeds reward fabricated stories and engagement-bait, making genuine shared experiences impossible to distinguish from fiction.
3. The Discovery Paradox: You cannot search for someone who experienced the same private moment without first exposing your own vulnerable, unverified thoughts to the public internet.

The Solution:
Echoed solves this by separating private submission from public discovery:
Zero-Exposure Journaling: Experiences stay 100% private in personal vaults upon submission.
100% Meta AI-Powered Verification: Meta Muse Spark Open acts as an validator, evaluating real convergence between independently written accounts while stripping PII.
Dual-Author Mutual Consent Wall: Even after a high-confidence match is detected, nothing is published automatically. Both authors independently review the real submissions and must explicitly click "Confirm Authenticity & Share Publicly." If either author says no, the submission remains strictly private in their personal vault forever.
Guaranteed Anonymity: When published, real identities are shielded behind persistent, poetic two-word aliases (e.g., MysticVoyager, CosmicDreamer), allowing people to find profound human connection with absolute privacy and data sovereignty.

## More from Meta Global AI Developer Hackathon

- [i-persons](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1938)
- [WordBlade](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1939)
- [Senti_One](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1940)
- [AssemblyGuard](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1942)
- [Wisdom Douglas](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1943)
- [HydroWatch](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1944)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1941. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt