-
-
Cover
Inspiration
The explosion of generative AI has made it trivial to create convincing images, videos, and voices of real people without their consent. In recent months, women, journalists, and public figures have been targeted by non-consensual deepfakes and impersonation campaigns that are almost impossible to stop once they spread. We wanted to build something that didn’t just detect fakes, but gave people a real, enforceable way to defend their identity online.
That became CreativeArmor: an active-defense system for the AI era.
What it does
CreativeArmor lets users protect their face and their content by anchoring it to a cryptographic identity.
A user takes a selfie or uploads an image. The app generates a face fingerprint, hashes the image, and embeds an invisible watermark. That proof is written to Solana as a permanent, tamper-proof record of authorship.
When a suspected deepfake is found, CreativeArmor compares it against the user’s registered identity. If the face matches but the image does not, the app generates a Certificate of Inauthenticity and produces an automated DMCA takedown notice with direct links to report the content on major platforms.
Instead of “I think this is fake,” users get “Here is cryptographic proof this is fake.”
How we built it
We built a modern Microsoft-style web app with a live camera scanner, animated face-scan overlays, and a glassmorphic dashboard.
The pipeline works like this:
- The camera captures a face and extracts a face embedding.
- The image is hashed and watermarked.
- A proof record (face hash, image hash, watermark hash, wallet address, timestamp) is written to Solana Devnet using
@solana/web3.js. - Gemini is used to generate legal-grade takedown notices.
- The UI displays verification status, Solana transaction hashes, and real-time protection results.
Solana acts as a decentralized identity and evidence layer, not a crypto app.
Challenges we ran into
- Designing a system that felt legally meaningful while still being demoable in a hackathon.
- Balancing real cryptographic verification with fast, reliable UX.
- Making blockchain feel invisible to users who just want protection, not wallets and tokens.
- Simulating real-world deepfake discovery without access to social media firehoses.
Accomplishments that we're proud of
- We built a working end-to-end flow from face capture to on-chain identity proof.
- We created a real Solana-based authorship registry instead of a mock database.
- We turned deepfake detection into an actionable legal workflow.
- We designed a clean, cinematic UI that makes AI safety feel tangible and human.
What we learned
We learned that AI safety is not just a technical problem, it is a trust problem. Detection alone isn’t enough — people need proof, provenance, and a way to act. We also learned how powerful blockchain can be when it’s used for public verification instead of speculation.
What's next for CreativeArmor
Next, we want to expand CreativeArmor into a full identity layer for the internet:
- Browser extensions to detect fakes in real time
- Platform integrations for automatic reporting
- Hardware integrations like Meta Ray-Ban glasses for on-device verification
- A global face registry that helps platforms distinguish real humans from synthetic ones
CreativeArmor’s goal is simple: give people their digital selves back.
Log in or sign up for Devpost to join the conversation.