SOLved

CAPTCHAs are failing to stop bots. We built the first advertisement-based security layer on Solana, turning outdated security into interactive visual puzzles built from real brand campaigns.

Comment

Inspiration

Our project began by staring at a wall, literally. When we first sat down to brainstorm, a teammate opened Google and whether due to a bad network, IP reputation, or just a broken laptop, we were hit with a Google Captcha bot check. As we spent time picking out crosswalks and traffic lights, the idea for a better Captcha was born.

We realized that while advertisements inside Captchas are still relatively new (and we did find one competitor), the concept had never been brought to Web3, specifically the Solana blockchain. Then a team member, who is also part of Superteam Georgia, managed to grab all the information as he could and we came to the conclusion that ad-based Captchas have never been on-chain before. Until now.

What it does

SOLved is an ad-based, on-chain CAPTCHA system built on Solana that provides robust bot protection while acting as a seamless monetization tool. It operates on three main pillars:

  • Anti-Bot Defence: A multi-layered CAPTCHA combining jigsaw puzzles, video challenges, browser fingerprinting, trajectory analysis, and memory-hard proof-of-work to stop bots while keeping humans flowing.
  • Built on Solana (On-Chain Micropayments): Every CAPTCHA solve triggers a SOL micropayment to the website owner. It features wallet-based sign-up, ed25519 signature auth, and instant transfers on the Solana devnet.
  • Behavioral Fingerprinting: Every CAPTCHA solve captures rich interaction data — mouse trajectories, drag patterns, and timing signals. Humans exhibit spatial understanding that bots lack, and predictive models leverage these patterns to reliably distinguish real users from automated scripts.

How we built it

Building SOLved started with intense brainstorming sessions among our team members to establish the core principles of the project. Once we agreed on the vision, we divided the workload into distinct parts, Web3 integration, image CAPTCHA, video CAPTCHA, and front-end design. Then, we built the entire system from the ground up.

To create the perfect balance, we had to change our entire CAPTCHA design exactly three times. Our goal was to build an interface that was:

  1. Computationally hard for AI and bots to solve.
  2. Simple and enjoyable for real human beings to interact with.

Challenges we ran into

We ran into 2 main challenges:

  • Time Constraints: We initially planned to write and implement the full smart contract to handle the SOL payouts during the hackathon, but due to time limits, we had to pivot to a simpler off-chain solution for our MVP.

  • Design Iteration & AI Evasion: Our biggest struggle was finding the perfect sweet spot for the CAPTCHAs. We wanted to make the puzzles easy and intuitive for humans, but computationally expensive and confusing for AI bots. We had to scrap our work and start over multiple times. For example, we deliberately added features like edge blurring to the video CAPTCHA to completely throw off edge-detection algorithms, which won't effect human perception.

Accomplishments that we're proud of

  • Building a Startup, Not Just a Project: Through our brainstorming and iterations, we realized SOLved has the legs to be a full-fledged startup.
  • Guaranteed Ad Engagement: We're proud to have built a model that actively solves "banner blindness." Because users must physically interact with the CAPTCHA to proceed, it ensures 100% engagement with the advertised brand, offering far higher value and recall than traditional ads.
  • Defeating Clawdbot: To put our security to the test, we ran an automated Clawdbot directly against our CAPTCHA system. It completely failed to solve the puzzle, validating the strength of our multi-layered defense and AI-evasion techniques!

What we learned

This project was a massive learning curve where we dove deep into many topics. We learned how to implement memory-hard Proof of Work (PoW) behind the scenes to computationally disincentivize bot networks, while simultaneously integrating SOL micropayments to instantly reward legitimate host traffic. Furthermore, we gained deep insights into behavior analysis using predictive models, understanding how backends capture and evaluate subtle interaction data, like cursor trajectories and timing signals, to reliably distinguish real human psychology from automated scripts.

What's next for SOLved

Our immediate next step is to tackle the challenge we couldn't finish in time: writing and integrating the actual Solana smart contract to fully decentralize the SOL generation and payout process for host websites. From there, we plan to take SOLved from a hackathon project to a fully launched Web3 startup!

Built With

Share this project:

Updates