Back to selected projects EDGE-BASED DEFENSIVE AI

womyn.network

A security-first, female-centric social media SPA exploring browser-side AI protection, privacy-aware posting, and safer community interactions.

Project Summary

womyn.network is an invite-only web app built as a hash-routed SPA with vanilla JavaScript, Parcel, Firebase Authentication, Cloud Firestore, TensorFlow.js and MobileNetV2. It includes signup and login flows, feed filters, text and image posts, threaded comments, profiles, follows, circles, notifications, settings, and privacy cloak image protection.

The public snippet documents early-stage browser-side-only adversarial imaging pipeline tests. A full demo is available upon request.

Technical Focus

  • Client-side defensive AI and image protection experiments.
  • Security-first social features for safer participation and moderation.
  • Firebase-backed authentication and data persistence.
  • Human-centred UX for a female-centric digital space.

Live Site & Funding

Ignite Fund Grant Winner: womyn.network has moved beyond thesis prototype stage with support from the Ignite Fund, recognising the project as a promising early-stage initiative worth developing further!

The womyn.network interest-gathering site is now officially live at womyn.network, using this grant-backed momentum to move the project toward an active public-facing platform.

I have also applied for further funding to support the next stage of development, community validation, and launch preparation. At this stage, the platform is being developed toward a text-based-only social network release while the AI elements, safety tooling, moderation systems, and privacy protections are refined responsibly.

The roadmap is intentionally staged: start with a focused social experience, keep safety and trust at the centre, then expand the defensive AI functionality as the development and marketing team grows.

Thesis Preview

Open in Google Drive

If the embedded preview does not load, the Google Drive button opens the thesis directly.

Demo Information

Open Link