I'm going to be very direct because I think you are on the right path: I don't think your main problem is the idea or even the pricing. I think the execution is currently killing the product.
I opened the site to try as a user and my experience went kinda like this:
The thing at the top immediately confused me. I couldn't tell what language I was looking at. Then I discovered that the "NO" button somehow switches the site to English. That alone made the product feel amateur. The site have a very strong half-finished vibe-coded feeling. I use AI heavily myself, so I'm not saying "AI = bad." The problem is that your target audience is technical. The people who understand why they would want an audit like this are also exactly the people most likely to think, "I could throw this into Claude/GPT myself." You need the implementation and the resulting analysis to demonstrate why they can't get the same result from a simple prompt.
Right now it does the opposite.
I generated a report. The presentation isn't great, but the bigger issue is that the analysis itself is technically wrong in several places. If you're auditing modern websites, you need a real JS-capable crawler/rendering step, structured extraction, separate analysis for different categories, carefully designed prompts/evaluators for each area, and an architecture that produces deterministic evidence for the claims in the report. A generic "send some HTML to an LLM and ask for an audit" approach isn't enough here.
Another big issue: Because of the vibecoded nature of the website, I assumed the locked parts of the report might simply be visually hidden because of how the rest of the application was implemented. I opened DevTools and, sure enough, the supposedly hidden content was already sent to the browser. Anyone remotely technical can retrieve the paid result without paying. Given your target audience, I suspect a lot of users will notice this.
Basically:
1. The UI makes me question the quality.
2. That makes me inspect the implementation.
3. The implementation confirms my suspicion.
4. That makes me distrust the audit.
5. Reading the audit then gives me technical reasons to distrust the actual product.
And that's a shame, because I actually think the basic idea has potential, the domain is great, and your pricing feels correct too for what you giving;
If I were you, I wouldn't throw the idea away. I'd stop adding features and rebuild the core audit engine until the report is good enough that a technical user reads it and thinks, "Okay, this caught things I wouldn't have caught with a couple of prompts.". Work on your technical skills first, seems like your marketing skills are WAY ahead of your technical skills right now, you should switch focus for a bit.
You're doing the right thing by building things, shipping them, failing, and learning from them. That's how you get better. But for this particular product, your audience has a very high technical bullshit detector, so the quality bar is substantially higher.
Good luck. Keep building.