Shipped my first SaaS in a month — AI meta tag generator with live social previews
A month ago I decided to finally ship something instead of endlessly planning. Today MetaBlast is live.
Paste any URL → AI reads the page → you get an SEO title, meta description, Open Graph and Twitter Card tags, with live previews of how the page looks on Google, X, Facebook and LinkedIn. Copy the HTML, paste into your <head>, done.
Pricing: free tier (3 generations daily, no signup), $9 one-time for 100 generations, $12/mo unlimited.
Stack: Next.js 16, Claude API, Supabase, Lemon Squeezy, Vercel. Built evenings and weekends with heavy use of AI pair programming — happy to share details about that workflow if anyone's curious.
Would love honest feedback — especially if something feels confusing or missing.
Thanks! The platform previews being the differentiator is exactly what I was betting on, so good to hear it lands.
On JS rendering: it reads the server-returned HTML, so fully client-rendered pages get limited context. Behind-auth pages (dashboards, Shopify admin) aren't really the use case though, those pages don't need SEO tags since crawlers can't reach them either. The tool is aimed at public-facing pages.
On the AI workflow: Claude handles the content analysis and tag generation, it gets the parsed page content and returns structured tags. The orchestration around it (fetching, parsing, validation, character limits, preview rendering) is plain TypeScript. I kept the AI surface small on purpose: one well-defined job, predictable costs, that's also what makes the free tier sustainable.
As for building it: I used Claude Code as a pair programmer throughout, but the architecture and product decisions were mine. Honestly a meta tag tool built with AI, about AI, reviewed by people who build AI tools — fitting.
This is a very unique idea,sometimes when I am looking for inspiration on what to build I end up in these periods of literally having no ideas. How do you think about such ideas?
For me it wasn't a flash of inspiration. I just wanted to ship my first public product after years of client work, so I was explicitly looking for something small enough to build in a month but real enough that people pay for it. Meta tags fit: clear scope, obvious output, easy to demo. I'd say don't wait for the perfect idea like i did for years :D. Pick a small real problem and ship and you learn more from one launch than from ten ideas in a notebook.
congrats on shipping. the $9 one-time tier is smart, most tools skip that middle option and lose people who don't want a subscription for occasional use.
does it handle SPAs okay or does it struggle when content is client-rendered?
Right now it fetches the raw HTML server-side, so it works best on anything server-rendered (Next, Nuxt, Astro, WordPress, plain HTML). That covers most pages people actually optimize (landings, blogs, e-commerce, docs).
Pure client-rendered SPAs are the weak spot, no JS execution means the AI only sees what's in the initial document. Headless rendering for those is on the roadmap. Though arguably if your marketing page is a fully client-rendered SPA, meta tags are the least of your SEO problems 😄
And thanks — the $9 tier exists because I hate forced subscriptions too.
Shipping in a month and getting a paid tier live on day one is the hard part, nicely done. The thing I'd flag as an operator: a meta tag generator is a utility, and utilities have no built-in reason to bring someone back next week. They fixed the tags once, so why return? The no-signup free tier makes that worse because you can't even email the people who tried it. I'd turn the one-time job into a recurring one: monitor a site's tags and ping me when a page is missing OG data or a title gets truncated after a deploy. That turns "I used it once" into "I keep it running," which is the difference between a tool and a business.
This is probably the most useful comment I’ve gotten since launching. You’re right, “fixed it once, why come back” is exactly the weak spot of the current model.
Monitoring is now at the top of my list: watch your pages, alert when OG data goes missing or a title gets truncated after a deploy. That also finally gives the subscription tier a reason to exist beyond higher limits.
And fair point on the no-signup free tier. I’ll rethink that trade-off.
Congrats on shipping. Going from planning to a live product is a bigger milestone than most people give it credit for.
One thing I like is that the value proposition is immediately understandable. Paste a URL, get tags, see previews. No learning curve required. That's surprisingly rare.
My first question would be whether users care more about generating tags or validating them. The live previews might actually be the stronger feature because they help people catch mistakes before publishing. If you notice users spending more time on previews than generation, there may be an opportunity to lean harder into the QA and testing side of the workflow.
Also, the one time credit pack is a nice touch. It gives occasional users a way to pay without feeling like they're signing up for another subscription.
Thanks! And honestly you might be onto something with the validation angle. People care about confidence before publishing, not just generating tags faster.
Right now I don't have per-feature analytics to see where users spend time, but that's worth adding. If previews turn out to be the main draw, a "validate existing site" mode (paste URL, get a report of what's broken or truncated) would be a natural next step.
The credit pack came from my own annoyance with subscription fatigue, glad it lands.
Meta tags feel like a small thing, but they quietly decide how much attention a page gets before anyone even clicks it.
One thing that often gets overlooked is how different platforms interpret the same metadata in slightly inconsistent ways, which can change click behavior.
Do users usually tweak the AI output, or do they trust it enough to ship it as-is?
Yeah, the platform differences are exactly why I added the side-by-side preview (Google, X, Facebook) so you see it before shipping.
Most people tweak the title a bit to match their voice, but the AI handles length and structure fine. Honestly I prefer it that way, blind trust in AI output is how you end up with generic meta tags anyway.
X is the worst offender btw, it ignores og:title half the time.
useful
Thanks! Anything you’d want it to handle that it doesn’t yet?
Interesting launch.
The thing I'd be careful with is that generating meta tags is the visible action, but that doesn't automatically mean it's the reason someone pays.
The risk is optimizing around usage before you're sure what decision, risk, or outcome the buyer is actually paying to avoid.
That sounds subtle, but it can change the entire conversion path.
I wouldn't make that call casually in a thread.
Good point, appreciate it.
My bet is people aren't really paying for the tags themselves — they're paying to not look broken when someone shares their link and the preview is missing or cut off.
That's why the live previews are the main thing, the HTML is almost a byproduct.
But honestly it's day one, I'll see what free users actually do before drawing conclusions.
That's the distinction I'd keep watching.
Sometimes what users say they're avoiding and what they actually pay to avoid end up being different things.
Day-one data is usually where that starts to become visible.
Hello Aryan Sinh
how was your day going?
One chilly Friday morning, I found myself scrolling through TikTok when I stumbled upon a broker that seemed too good to be true. The scammers claimed they were offering pool investments for Bitcoin investors, allowing individuals to choose from various packages to invest. They provided videos of real trades and impressive profit-and-loss (PNL) screenshots, showcasing their success and enticing me to join in. Their polished presentations and persuasive tactics made it easy to believe that I was on the brink of significant earnings. Eager to replicate their success, I invested $96,300, convinced that I was making a smart financial decision; it didn’t take long for the reality to set in. I tried to reach the broker for updates, but communication slowed down. My messages went unanswered, and the once vibrant community I had joined became eerily silent. It soon dawned on me that I was dealing with a fake broker. The scammers vanished, taking my funds with them, leaving me feeling anxious, helpless, and utterly betrayed. I reached out to friends for advice. One of them recommended TechY Force Cyber Retrieval, a team specializing in recovering lost funds from scams. Skeptical but hopeful, I decided to give TechY Force Cyber Retrieval a try. That turned out to be the best decision I could have made. Their team acted swiftly, thoroughly investigating my situation. TechY Force Cyber Retrieval was professional, empathetic, and dedicated to helping me recover my money. Throughout the process, TechY Force Cyber Retrieval kept me informed, providing updates and reassurance when I needed it most. They meticulously analyzed the details of my case and employed effective strategies to track down my lost funds. TechY Force Cyber Retrieval not only helped me recover my lost $96,300 but also the promised profits that had initially lured me in. I was absolutely thrilled and incredibly relieved to see my funds returned. This taught me a valuable lesson about the importance of being cautious in the world of online bitcoin trading. I realized that while scams can happen to anyone, there is hope for recovery with the right support. If you’ve ever been scammed or are struggling to withdraw your investments from a broker, I highly recommend TechY Force Cyber Retrieval. They are experts in fund recovery and can assist you in regaining what you've lost. Their approach and dedication to client success are truly unparalleled. Don’t let a scam define your financial future; seek help from TechY Force Cyber Retrieval and take action. You deserve to reclaim your hard-earned money.
WhatsApp +.1.5.6.1.7.2.6.3.6.9.7
Telegram +.1.5.6.1.7.2.6.3.6.9.7