How to Launch on Product Hunt in 2026 (With AI)

Half the Product Hunt advice still floating around was written for a site that stopped existing four years ago. The celebrity hunter who auto-tweeted your launch, the leaderboard that rose and fell on raw upvote count, the coordinated 9am spike followed by radio silence: those mechanics are gone, and playbooks built around them now produce launches that stall before lunch.

What works in 2026 is almost the opposite in spirit. Product Hunt scores launches on points rather than upvotes, weighs genuine discussion heavily, and strips out votes it judges inauthentic. The winning move is no longer mobilizing a vote block on the morning of launch day. It’s arriving better prepared than the thirty other products fighting for the same homepage. That is also where AI earns its place: tagline variants, gallery drafts, a demo script, a reply plan, all built in advance on a founder’s budget.

A note on who I am: I’m James Henderson, and I run this site solo on top of a fairly obsessive automation stack. I have never personally taken a product to #1 on PH, so treat this as a synthesis rather than a war story. It’s assembled from Product Hunt’s own help documentation, from the trackers that watch the leaderboard daily, and from the post-mortems indie makers publish every week. When something is my opinion rather than documented fact, I’ll flag it.

What Actually Changed: Points Beat Upvotes

The most persistent misconception in launch circles, even among people who follow the site closely, is that upvotes are the score. They’re an input. Product Hunt’s points documentation is blunt about it: every launch gets assigned points, points determine homepage ranking, and one upvote does not always equal one point. Featured status matters, overall community engagement matters, and the system also looks at how authentic the incoming votes appear.

Comments carry real weight in that mix. A threaded discussion from members with account history and varied locations outperforms a burst of drive-by upvotes from brand-new profiles. Nobody outside Product Hunt knows the exact formula, and the company says the silence is deliberate to keep the system from being gamed. So any guide promising hourly upvote targets is guessing, and I’d rather you know that upfront.

The published warnings matter more than the hidden formula. Product Hunt’s page on asking your community to upvote opens with two words: “Please don’t.” Their reasoning is that people should upvote things they genuinely find interesting, not because they were peer pressured, and asking for or incentivizing upvotes can get a product demoted in the rankings or removed from the homepage entirely. Enforcement is real as well: algorithms that watch for unusual voting patterns, community reports, and manual review by staff, with non-genuine votes removed along the way.

Here’s the practical version, uncomfortable but simple. Upvote pods, paid vote services, and “you upvote mine, I’ll upvote yours” deals are the fastest way to lose a launch. The votes get stripped, and products that looked like they were climbing at 9am have slid off the homepage by noon; launch trackers have watched that exact pattern repeat all year. You’ll get much further telling people your product is live and inviting them into the conversation. That part Product Hunt explicitly allows, and it dovetails with how the ranking system works now.

Before Anything Else: Can PH Even Feature You?

Featuring is curated, not automatic. Product Hunt’s featuring guidelines say the team looks for usefulness, novelty, high craft, and creativity, and that most featured products spike in one or two of those rather than checking all four boxes. That’s a lower bar than it sounds, though it’s also a different one than most founders expect.

The exclusion list deserves a slow read, because solo founders tend to build precisely the things on it. Product Hunt generally does not feature waitlisted products, directories, templates, boilerplates, podcasts, courses, reports, events, books, services, deal platforms, e-commerce storefronts, or hardware without a meaningful digital component. Vaporware and anything that reads as monetization-first gets rejected too. If what you built is a really good Notion template, don’t spend a launch day that can’t rank; sell it where template buyers already shop and put launch energy there instead.

Relaunches have their own rule: at least six months must pass between posts for the same product, and the update needs to be significant. A new mobile app or a redesign that ships genuinely new functionality qualifies. A fresh coat of paint or a pricing change doesn’t. Worth checking before you book the day, rather than discovering it after.

This is a legitimately good AI task, incidentally. Paste the featuring guidelines and the exclusion list into a capable LLM along with your product page copy, then make it argue against you: does this qualify, where are the weak spots, which exclusion bucket is closest? In my experience the model is sharper at spotting obvious disqualifiers than at taste judgments. As a ten-minute preflight pass, though, it catches exactly what excited founders miss on their own products.

Four Weeks Out: Warm the Room Without Begging

The runway is where launches actually get decided, and almost none of it involves votes. Four moves matter:

• Build a launch list: existing users, beta testers, newsletter subscribers, anyone who ever replied to you.
• Tell the build story in public: the problem worth solving, and the decisions that didn’t work.
• Rehearse the day itself: the maker comment, the reply plan, and the follow-up emails, all drafted early.
• Get checkout and onboarding ready, so the traffic lands on something solid.

That launch list should live in email, because email is the only channel you actually control when the day is over. Founders lean toward ConvertKit for exactly this job, which is a reasonable pick if you plan to keep marketing to the list after launch week. My own newsletter runs on Buttondown, which is deliberately less feature-heavy. That’s a taste call, not a verdict on Kit; the honest tradeoff is that ConvertKit’s power shows up as more settings and more pricing tiers to wade through.

AI pulls its weight in the runway period in one specific way: drafting the sequence rather than sending it. Have an LLM turn your rough notes into a set of build-in-public posts, roughly one a week, then rewrite each until it sounds like something you’d actually say to a person. Scheduling, meanwhile, belongs to real tools. My own publishing pipeline here runs on Make and n8n, and I’ve automated enough of this site to have opinions. Those tools schedule and draft impressively well. They cannot be a person in somebody’s replies, and both the audience and the platform can tell.

Assets: Build to Spec, Not to Hope

Product Hunt publishes its asset specs openly, so build to them instead of resizing something at midnight before the day. The working set:

Asset	Spec	Notes
Thumbnail	Square, 240×240 recommended, under 3MB	Static or GIF; animated thumbnails are common but not required
Gallery images	Minimum two, 1270×760 recommended	Reorderable, animated GIFs accepted
Video	Optional, YouTube links only	A 60 to 90 second demo of the real product in use
Tagline	60 characters max	Product Hunt asks for no gimmicks or hype language here
Description	500 characters max	What it does, key features, who it is for

Two of these decide more than the others. The tagline is your entire pitch for a stranger scanning thirty launches, so generate a pile of variants; fifteen is a fine floor. Draft them with an LLM by all means, then read them out loud to people who don’t work with you, and keep the one that produces a question rather than a polite nod.

Gallery images have quietly standardized on Canva templates, which is why so many founder galleries look properly finished: Canva has become the default for non-designers building launch assets fast. If you already pay for Adobe, use that instead. Nobody evaluates your tool stack, only the output, and nobody can tell which app made a good image.

The video is the long-lead-time asset with a real payoff. Launch trackers report that roughly half of Product of the Day winners include one, which matches what I see scrolling the homepage: products showing a real interface in motion read as more credible than polished brand films. A sixty to ninety second screen demo, uploaded to YouTube early with a proper title, also keeps earning views long after launch day is over.

You don’t need a production budget for that video, either. Any screen recorder works, and if you don’t have a favorite yet, Loom is the sensible default for a founder-facing product walkthrough. Script it with an LLM if that helps you start, then throw the script away and say it like a person; viewers forgive rough edges, not stiffness.

The highest-return paragraph of the whole month is the maker’s first comment. By Product Hunt’s own reporting, roughly 70 percent of products that went on to become Product of the Day, Week, or Month launched with a first comment written by the maker. Cover why you built this and what it still doesn’t do. Then ask for one specific kind of feedback, because a specific ask invites specific replies, which is precisely what the ranking system rewards. Write it days before launch, not at 1am while the page is already live.

Launch Day: The Sequence That Works

Submit at 12:01 AM Pacific on Product Hunt. The daily leaderboard resets at midnight Pacific time, so anything posted later runs the same race on a shorter lap. The exception launch trackers mention: if your entire audience sits on the US East Coast with no international users, a later slot can sometimes work. Almost nobody actually has that constraint, so plan for the full day of runway.

When the page goes live, post the maker comment immediately, then work the thread. Reply to every comment you get, answer skeptics seriously instead of defensively, and stagger yourself (or your support person) so replies keep arriving across the whole day instead of clustering in the morning. Products that keep the discussion alive are the ones the ranking system keeps pushing.

Your own channels need exactly one message, and it shouldn’t be about voting. Say the product is live, tell why in two sentences, and invite questions. In the Discord community I run for this site, posts that ask the room a question reliably draw replies, while posts that ask people to go do something get ignored. It’s not a moral insight, just an observed one, and PH’s moderation punishes the second version anyway.

Last opinion of launch day: stop refreshing the leaderboard. Early ranking jitter is normal and means nothing, and watching it all day wrecks the focus your thread replies actually need. Check the board twice, work the comments ten times, email your list once.

Where AI Helps (and Where It Backfires)

Being specific about what AI is worth on a launch, from watching how these runs actually go:

• Tagline and copy variants: the strongest use, hands down.
• Gallery drafts and thumbnails: solid, especially inside template tools that already do the layout for you.
• First-comment drafts: a good starting skeleton that you then humanize heavily.
• Competitor launch research: reading old launches and post-mortems is much faster with an LLM summarizing.
• Post-launch recap work, including threads and the newsletter follow-up: where AI genuinely shines, since it repurposes what already happened.

Where it backfires: engagement. Bot-written comments get caught by voters and by Product Hunt’s enforcement alike, and AI replies that don’t match reality read as fluff to the people deciding whether your product deserves attention. My own content operation is heavily automated, which is exactly why I’m confident saying this: automation schedules the work, but it cannot be a person in the thread. I hand-edit every article this site publishes for that reason, and every launch reply deserves the same fingerprints.

The Week After Is Where Launches Are Won

Set expectations with real numbers first. Launch trackers consistently report that B2B tools should expect 50 to 300 signups from a featured launch, while consumer products that place high on the board can see 500 to 2,000. A few hundred quality visitors for a solo founder is a genuinely good day, especially with a working signup funnel behind the product rather than a dead end.

Day two is for follow-through: email everyone who came, onboard the signups, and fix whatever they stumbled over. Post a changelog that replies to the launch thread itself, because the people who commented will come back, and a returning visitor converts at a multiple of what their first visit ever would.

Launch traffic without a capture plan is just a warm gust. If you’re in that funnel-building mood, my first 100 customers playbook is the natural next read; it covers turning attention into a pipeline that survives launch week. And set up checkout before the launch rather than after, not during it; an older post here compares Gumroad, Lemon Squeezy, and Payhip for exactly that wiring.

Key Takeaways

• Rankings run on points, not raw upvotes, and comments plus authentic engagement tilt the score.
• Upvote pods get their votes stripped, and asking for upvotes can demote or remove a launch entirely.
• Check featuring eligibility first, because the exclusion list covers a lot of what solo founders build.
• Start the runway four weeks out: launch list in email, build story in public, replies rehearsed.
• Build assets to spec: 240×240 thumbnail, two gallery shots, a 60 to 90 second demo video, one 60-character tagline.
• Submit at 12:01 AM Pacific and post the maker comment the minute the page is live.
• Use AI for prep and repurposing, then keep every launch-day reply human.
• The week after decides the value: email the list, onboard fast, fix out in the open.

Ready to Build the Rest of the Stack?

A launch is one afternoon. The business is everything you make repeatable so the next one is easier, and that’s mostly systems: pipelines, templates, checkout, onboarding. If you want a structured shortcut for setting up an AI-assisted founder operation, one course that fits this site’s audience is The AI Blueprint, which covers the automation and workflow side of running a one-person company. That link is an affiliate one; the short version of what that means sits right below.

Affiliate and methodology disclosure: some links in this post are sponsored or affiliate links, which cost you nothing extra and help keep the site running. This guide was written in late September 2026 from Product Hunt’s published help documentation, launch trackers, and public maker post-mortems, with figures attributed where they come from those sources. I don’t claim a hands-on launch of my own; where I name tools I actually run, like the Make and n8n pipelines behind this WordPress site, or my Buttondown newsletter, that comes from real use, and everything else is research plus opinion.