
Cue
Curated Awwwards-tier UI components, reusable in any stack
Hey IH,
Alok here — solo founder from India, first real launch. This community has been my 2 AM read all summer, so figured I'd give back with an honest debrief.
TL;DR
- Product: Cue (cuedesign.space) — hand-curated UI component library with AI prompts for Bolt, v0, Cursor, Framer AI, Claude, ChatGPT
- Solo built, no VC, no ads, no growth hire
- Launched end of August 2026
- First paying user: 15 minutes after go-live
- 16 days in: $900+ revenue, ~10 paying founding members
- 40+ founding slots remaining ($99 lifetime, capped at 50, then $249)
The build
I spent all summer talking to designers and indie devs about their AI-tool workflow before writing a line of code. Same pattern every time — AI builders (Bolt, v0, Cursor) can generate anything, but the output quality is entirely a function of the REFERENCE fed in. Generic prompt → 2018 gradient hero. Reference anchored in a specific Awwwards-tier interaction → something you can actually ship.
That gap became the product. Cue is that anchoring work packaged — every component is hand-picked from an Awwwards Site of the Day, a Behance-featured interaction, or a production reference I rated best-of-class. Nothing machine-generated to fill the grid.
What worked
-
Building in public on X, but for accountability not followers.
Posted the ugly WIPs. Wrote about specific problems — Clerk auth quirks, Dodo webhook race conditions, Supabase RLS pain. Built a small audience of ~500 people who actually cared. When I launched, they were the first click. -
Shipping embarrassingly early.
V1 landing page was rough. But signup was clean, Dodo payments worked, and the actual value (curated components + AI prompts) was there. Every polish decision I skipped in July saved a week I could spend curating more components. -
Lifetime pricing over subscription.
$99 lifetime for the first 50 founding members. Not a subscription. After founding sells out: $249. Made the founding tier a public promise — no future price change ever applies to founding members. Founding buyers converted faster than any discount timer would have moved them. -
Distribution stack, not just Product Hunt.
PH is a 24-hour spike. Real growth came from:
- Directory submissions (SEO backlinks that compound over months)
- Weekly component drops via newsletter
- Cold DMs to designers who commented "wish this existed" on my WIPs
- ScrollLaunch, IndieHackers, HN threads answering questions honestly
- llms.txt + llms-full.txt so AI answer engines (ChatGPT, Claude, Perplexity) can cite Cue when users ask about component libraries
- Founding-member personal handoff.
Anyone who upvotes on Product Hunt + DMs me on X gets one premium component unlocked free, within an hour, by me personally. This one move drove more conversion than any automated funnel.
What I got wrong
-
Got publicly called out for uncredited inspiration.
Built one component clearly inspired by a specific designer's work without crediting them. Got called out on X. Removed it same day, apologized, and I'm adding a Credits field to every component schema going forward. Painful public lesson but the right one to learn early. -
Underestimated distribution work.
I thought PH was "the launch." PH is a spike. The compounding work — directory submissions, SEO content, showing up in AI answer engines — is the actual plan, and I'm still catching up on that. -
Over-polished before launch, then had to un-polish under pressure.
Wasted maybe 4 weeks on animation experiments (video cursor trails, glass overlays) that I later had to remove because they hurt conversion and looked AI-generated. If I could redo the summer, I'd ship 2 months earlier and iterate live. -
Initial pricing had a monthly tier that made no sense.
$49/month tier ran alongside $99 lifetime. Broken ladder — a rational buyer would never pick monthly when lifetime is cheaper. Killed it after launch week. Lesson: pricing tiers need to reinforce each other, not compete.
Stack (for the tech-curious)
- React + Vite on Vercel
- Clerk for auth (email OTP + Google OAuth)
- Dodo Payments as Merchant of Record — handles local currency + tax globally, way less friction than Stripe for an India-based founder
- Supabase Postgres with Row-Level Security
- Cloudflare R2 for component preview media
- Resend for drop emails (batched via a Supabase edge function to respect rate limits)
- PostHog for analytics (cookie-free)
What I'd love feedback on
- How would you have structured the founding-tier vs standard pricing?
- Anyone here run distribution for a curation product — what channels actually compound?
- Is "AI builder taste layer" the right positioning, or should I lead with something more concrete?
Site: https://cuedesign.space
Full technical + product context: https://cuedesign.space/llms-full.txt
Twitter: @Alok619308
Happy to answer anything specific — pricing, tech, distribution, whatever. Would love brutal honest feedback from this community especially.
— Alok
Third post here. Third month in. Sharing what happened today
because it's genuinely a data point I don't want to lose.
Ran a weekly-drop thread on X earlier today — same format
I've been posting every 5-7 days for 6 weeks. Showed off the
3 new components I dropped this week, one screenshot each,
one line each.
Result today:
• 15k+ impressions (my usual is 1-2k)
• 600+ unique visitors to cuedesign.space
• 200 total signups now on Cue (was ~120 yesterday)
Nothing viral about the post itself. What did seem to compound:
What worked:
1. Consistent daily+ presence for 6 weeks. Not one post did
this — it's the trust built by showing up. First 5 weeks
of posts were flat. Week 6 they started ranking.
2. Component videos are the hook. Screenshots don't stop
scroll — motion does. Every post now has a looping mp4.
3. Reply to EVERY comment within 15 min. X algorithm
rewards active conversations, and it's the difference
between a 1k post and a 15k one.
What didn't work (yet):
1. Conversion from visitor → signup is only ~33%. Should be
40-50% for a taste-driven product. Hero copy or the
paywall UX is losing the other 200.
2. Discord invite in bio → 30-40 joins from ~200 signups.
Fine but not amazing.
3. Zero paying customers from this specific spike. Cue+
sales are slower than signups by weeks — buyers browse
for a while before committing.
Feedback wanted:
1. Anyone else run a "consistent weekly drop" cadence — when
did compounding actually kick in for you? Week 6 feels
early but real, want to know if this is a false peak.
2. 33% signup rate on 600 visitors — what would you look
at first to move it to 50%? (Hero copy? Paywall?
Component previews below fold?)
3. How long does browser → buyer lag look for you on a
$99 one-time product? I'm assuming 2-3 weeks but
flying blind.
Reading every comment.
1 Like
Comment
Second post here. Last week's launch went okay — this one's about how the 3 months actually went. Nothing polished. Just what happened.
Month 0 — obsession phase:
Not "kinda interested" — obsessed. Every free hour spent watching dev breakdowns, dissecting how Awwwards sites got built, opening Figma files, reverse-engineering React source, reading component library READMEs at 2am. Six weeks of consuming before writing a single line of code.
If you're skipping this to "just start shipping" — you're probably building the wrong thing. Mental library first, code second.
Month 1 — talking to people, not coding:
Before building anything, Dm 30+ designers and devs asking:
- "What component library do you actually use?"
- "What's broken about it?"
- "What made you give up and just build from scratch?"
Their answers became the loopholes I built around:
- "Everything looks the same" → curated Awwwards-only
- "No AI prompts, just code" → added prompts tuned for Cursor / v0 / Bolt
- "Nothing is production-ready" → tested every component in a real project before adding it
Month 2 — building + shipping tiny:
Started slow. Dropped 5-10 components per week on X. Consistency was the only lever that mattered. Not viral posts — just showing up daily. Small screenshots, small demos, small threads.
X follower growth was slow. But the RIGHT people started following — designers, indie devs, people building for AI coding tools.
Month 3 — launch day:
Site went live. Posted on X. Told ~800 followers.
First paid customer in 15 minutes.
Sat staring at the Dodo notification for 20 minutes, refusing to believe it. Then DMed the buyer to thank them personally.
Next few days: 7-10 more paid users. Not viral, not exploding — just enough to prove the thesis wasn't wrong. Small but real.
What actually worked (in order):
1. Obsession before coding — 6 weeks of consuming before shipping saved 3 months of building the wrong thing
2. User research DMs — 30 conversations gave the exact positioning
3. Daily X consistency — not for followers, for accountability + audience
4. Shipping even when scared — the launch tweet was terrifying; hit send anyway
What I got wrong (that I don't see other founders talk about):
1. Cheaped out on tools early — spent 2 weeks rebuilding what a $30/mo tool would've done in 20 min. That "savings" cost me a paying customer's worth of time.
2. Isolated too long — didn't join founder communities (like this one) till month 3. Should've been here month 1 for feedback loops.
3. Waited for a "perfect" landing page — rewrote it 12 times before launch, then rewrote it 4 more times AFTER launch. Would've learned faster with an ugly v1 shipped week 1.
Site: cuedesign.space
X: @Alok619308
Reading every comment.
2 Likes
5 Comments
5 Comments
-
1The first 8–11 sales validate willingness to pay, but the next signal seems more important: are those buyers actually using the components in projects and coming back for more?
-
1Basically, Cue offers lifetime access, and people who join become Founding Members. I regularly check in with them to understand how things are going and what could be improved, because user feedback is one of the most important parts of building Cue. So far, the response has been really positive. Some members are already using Cue to build better client projects and charge more for their work, while others are using it as part of their agency workflow.
-
1
That’s stronger evidence than the sales alone. I’d be interested in digging into what those users are actually using repeatedly. If you’re open to it, what’s the best email to reach you on?
-
-
Hey IH,
Been lurking for months. First post.
The problem:
Every side project I built, I lost 4+ hours hunting Awwwards / X / random blogs for the ONE hero animation, scroll effect, or preloader that was actually usable. Awwwards has thousands of stunning UIs, but 95% are portfolio demos you can't ship. The other 5% are buried across 20 tabs.
I got tired of it. So I built the thing I wanted to exist.
The solution — Cue (cuedesign.space):
A curated library of 140+ UI components. Every single one hand-picked by me from Awwwards Site of the Day winners, viral X interactions, and production apps that shipped it best. Every component ships with an AI prompt tuned across ~50 iterations — so when you paste it into Cursor / v0 / Bolt / Claude, you land on the version worth shipping, not iteration one.
The core idea: someone else's taste doing the filtering. You browse instead of scroll. You ship instead of hunt.
Early traction (3 months in):
- 500+ signed up combined (waitlist + accounts)
- First paying customers converted within 15min after launch
- Building in public daily on X (@Alok619308)
- Users started DMing "can I get JUST this one component" — that's driving my next experiment
- Bootstrapped, self-funded, solo, India-based
Site: cuedesign.space
X: @Alok619308
Any thoughts, roasts, or 'you're missing X' — reading every comment.
1 Like
Comment
About
I kept losing hours hunting for the ONE usable hero / scroll / preloader across Awwwards, X, and 20 tabs of open browsers. Cue is the taste layer — someone else curates the good ones so devs and AI builders can just ship


1 Comment
$900+ and ~10 buyers gives you real signal. Do those buyers describe Cue’s value as better references, better AI output, or something else entirely?