Cap is a fast, lightweight, and modern open-source CAPTCHA alternative based on SHA-256 proof-of-work (PoW). It’s 250x smaller than hCaptcha, privacy-focused, fully customizable, and easy to self-host. Cap helps prevent spam and abuse without tracking users, making it ideal for privacy-conscious developers. Built for speed and simplicity, Cap is perfect for modern web apps, forms, and APIs that need secure, lightweight human verification.
I was very pessimistic about this product—at some point, I thought it wouldn’t help at all. Now I’m happy. It just stopped the spam (registration attempts and contact form submissions) Good job!
Love this approach! 💡 Using SHA-256 proof-of-work as a CAPTCHA alternative is such a smart, elegant solution — especially in an era where user privacy and page speed really matter. 🔒⚡️
Curious how it performs in real-world bot scenarios — any benchmarks or early adopter feedback?
250x smaller than hCaptcha is huge for web performance. Can we modify PoW parameters per use case, like stricter thresholds for login vs. comment forms?
@desmond_ren1 yes! you can fully adjust the difficulty
Report
I'm a bit concerned about it's effectiveness. Prove me wrong, I'd be happy if this works as good as the others.
First, this does not verify if I'm a human, but if I have enough computational resources. A similar system was developed (Hashcash) which is not really used in popular email clients. In my opinion, it works for Bitcoin for the same reason it didn't work work email: it doesn't verify if you're a human, it just verifies your computational resources.
This raises some questions. What if someone is browsing my site from an old computer? The verification will take a lot longer and possibly use all the resources that device has for minutes.
What happens to botnets? While tracking-based captchas have a chance to combat them, it doesn't really matter if hacker guy has to do some PoW on the botnet computers.
Thanks to Bitcoin, we also have really efficient sha256 ASICs - computers that only solve sha256, but they do it really efficiently. If a verification take 2 seconds on a CPU, then it will take milliseconds on an ASIC. So with just one ASIC, I'm able to essentially break any website.
Right now I think this captcha is MUCH better than not using any captcha - but I don't think it is better that the tracking based captchas. I'd be the happiest if this could work, so please prove me wrong if I didn't get it right. I also think it is really important to have experiments like this, I really support the direction.
botnets can't really solve the captcha in a reasonable amount of time since they're usually very low-powered devices such as security cameras or routers
One question remains. So the PoW algo. Makes it prohibitive for bots to cheat the captcha. How much cost does it penalize bots with? is this CPU watts or? Like 0.1$ in computation?
And do you think this is the future of captcha? Mine some "bitcoins", to fight bots?
@tr3 It doesnt say anything about how much cost is incurred on the bots. Some transparency around this would be gold. Also Does it increase the difficulty of the challenge on repeated attempts? If not, may this be a future feature? The research paper you linked to describes a password cracking functionality. Is cap used for this purpose? That raises ethical dilemma, and End user legal dilemmas. Also how does it perform on mobile devices with limited computational resources and battery? Some transparency around this would also be great. Btw. I think this project is really cool! So that's why I'm curious. 😸 I also asked pplx regarding the research paper. And it echoed some of the questions: https://www.perplexity.ai/search/what-are-some-of-the-pros-and-uy_EbHZ1TsSlpMGSkFVTXw#0
Cap
thank you!