Beacon - Women's Hack Night

Beacon lets students report harassment confidentially. If others report the same person, Beacon connects their reports so they can go to Title IX together instead of alone.

Comment

Here's a full draft. Check the two items in [brackets] against what you actually built.

Inspiration

Earlier today, while news of a campus shooting spread, many students first heard about it through anonymous Yik Yak posts rather than through official channels. It was a clear example of something women have always done: when institutions are slow to respond, we warn each other.

Women have long relied on "whisper networks" to share who and what to avoid. In one national survey, 81% of women reported experiencing sexual harassment in their lifetime, yet most incidents are never formally reported. Coming forward alone is intimidating, especially when the other person holds more power, as Olive's story in The Love Hypothesis shows. Many women stay silent because they believe no one will take their word alone seriously.

We built Beacon to turn those private warnings into something that can actually lead to action.

What it does

Beacon is a confidential reporting platform for college students, starting at UC Berkeley.

  1. Tell it: Users describe what happened in a private chat. Beacon's AI asks a few gentle follow-up questions and turns the account into a clear, structured report under an anonymous alias.
  2. Seal it: Each report is timestamped and given a SHA-256 fingerprint the moment it's filed, so any later edit would be detectable. No one else can see it, including other reporters.
  3. Match it: When another student independently reports the same person or location, Beacon links the reports: "2 people, one pattern." Neither reporter learns the other's identity.
  4. Act on it: Beacon generates a case file with a combined timeline, the Title IX pathway, and Berkeley resources such as PATH to Care and OPHD. The case moves forward only when every reporter consents, so it reaches the Title IX office as one corroborated report instead of several isolated ones.

Nothing is ever posted publicly. Survivors stay in control at every step.

How we built it

  • Frontend: Next.js (App Router), TypeScript and Tailwind CSS, designed mobile-first as a progressive web app (PWA).
  • AI intake: the Claude API powers a trauma-informed intake assistant that asks follow-up questions and extracts structured fields (time, location, category, description of the person involved, severity).
  • Matching: reports are linked when they share a normalized identifier, such as a social media handle, [and Claude compares descriptions, location and timeframe to flag likely matches].
  • Integrity: SHA-256 hashing for each report, plus a combined case fingerprint generated from all of the case's report hashes.
  • Storage: [a lightweight local data store / Supabase].
  • Development: built in roughly two hours using Claude Code.

Challenges we ran into

  • Balancing safety and fairness. A tool for warning others can easily be misused to accuse someone publicly. We designed Beacon so reports are never public, matches stay anonymous, and nothing reaches Title IX without every reporter's consent.
  • Writing a sensitive AI intake. The assistant had to collect useful details without sounding clinical, pushy or judgmental.
  • Scope under time pressure. With two hours, we cut features like a live safety map and PDF export to focus on the core flow: reporting, matching and building a case.
  • Team coordination. We worked in different editors, so we had to set clear file ownership and keep the repository in sync.

Accomplishments that we're proud of

  • A working end-to-end flow, from a private report to a corroborated, consent-based Title IX case file.
  • A matching experience that shows survivors they are not alone, without revealing anyone's identity.
  • Tamper-evident records that support both survivors and due process.
  • Real, actionable Berkeley resources built directly into the case file.

What we learned

  • Building for survivors requires designing around trust, consent and privacy first, with features second.
  • Many people hesitate to report not because they lack options, but because they feel isolated. Showing that others share their experience can matter as much as the reporting tool itself.
  • Scoping tightly and building with AI coding tools let us ship a meaningful prototype in a very short time.

What's next for Beacon

  • Campus partnerships: pilot Beacon with UC Berkeley's Title IX office (OPHD) and ASUC, then expand to other UC campuses.
  • Community safety map: an anonymized heatmap of reported areas, with location-based alerts when users enter a hotspot.
  • Stronger privacy: end-to-end encryption and verified student accounts to reduce false reports.
  • Beyond campus: versions for workplaces, venues and nightlife, plus a native mobile app.

Built With

Share this project:

Updates

Submission history