Product Hunt and Hacker News Launch Playbook for an AI Side Project

You've built the thing, polished the page, and now there's a button called Submit. A Product Hunt Hacker News launch feels like a moment, but it is mostly a pile of small tasks that go wrong if you do them at 11pm the night before. This post is a process: what each platform's own rules say, what to prepare in the weeks before, what to do on the day, and what to do when nobody shows up.

I'm not going to promise a number of upvotes, visitors or customers. Nobody can, and any guide that does is guessing. What follows is process only, with platform rules read on 2026-10-09. Rules change, so re-read the linked pages the week you launch.

What do the platforms actually say?

Hacker News

Two pages matter: the Show HN rules and the site guidelines. Both were read on 2026-10-09 through a summarising fetch tool, so treat my wording as a paraphrase and check the originals.

From the Show HN page:

• It's for something people can run or hold, that you worked on personally and are around to discuss.
• Landing pages, sign-up pages, newsletters, blog posts, lists and fundraisers do not qualify.
• Work that isn't ready for users yet should wait.
• It should be easy to try, ideally without signups or email gates.
• The title must begin with "Show HN".
• Don't ask friends to upvote or comment.
• The page also says not to post quickly generated one-offs. If your AI project is a thin wrapper made in an afternoon, read that line twice.

From the guidelines:

• Don't use HN primarily for promotion.
• Don't solicit upvotes, comments or submissions.
• Don't use generated or AI-edited text in posts, and don't automate posting.
• Don't delete and repost.
• Titles: no caps or exclamation marks for emphasis, don't editorialise, drop the site name.

That "no generated or AI-edited text" rule affects how you use AI in this process. Draft your Show HN post and your replies yourself.

Product Hunt

The launch page, also read on 2026-10-09 through a summarising tool, says:

• Launches come from personal accounts; company accounts are not allowed.
• You can't ask people directly to upvote. Ask them to visit and comment.
• 12:01 am Pacific Time is the recommended time for planned launches, and the best day is the one you are most prepared for.
• You don't need a third-party hunter; you can hunt your own product.
• You can launch again after significant new iterations.
• The page points to a fuller launch guide and community guidelines that I could not read through the tool. It does not list required assets in the part I saw, so check that guide for the current asset list.

Notice what the two platforms share: they want a real product people can try, from a real person who will answer questions, and they dislike organised voting.

Which one first, or both?

They suit different projects.

If your tool needs an account before anyone can see anything, fix that before Show HN. A public demo, a sample output page or a no-login trial is worth more than any copywriting. If you're not sure either audience fits, skip both and post to a small community where your specific user hangs out. A launch is not mandatory.

Four weeks out: get the foundations ready

Launch traffic exposes every weakness at once. Work through these in order.

1. A page that explains the product. See the landing page prompts for solo founders for a structure and a review checklist.
2. Email that works. Signup and reset emails must arrive. Transactional email compared covers free-tier caps, which a launch-day spike can hit.
3. Abuse and cost controls. Free AI usage plus public attention is how bills spike. Put limits in first: rate limiting a free tier AI app.
4. Know your margin. If every free user costs real money, calculate it with the gross margin calculator per user.
5. A server you've secured. If you self-host, run through the VPS hardening checklist.
6. Analytics set up so you can see where visitors drop off. Without it, the launch teaches you nothing.
7. Privacy and data handling written down, in plain words, linked from the page.

One week out: assets and rehearsal

• A 30-to-60 second demo recording of the real product doing a real task. No mock-ups.
• Three to five screenshots at readable size.
• A one-line description in plain words, with no hype terms.
• A maker comment / first comment draft: who you are, why you built it, how it works at a high level, what it can't do, and what feedback you want. Write it yourself.
• A friend who has never seen the product tries it from the public URL on a phone, while you watch and say nothing. Write down every place they hesitate.
• A kill switch: one environment variable or config flag that disables free usage or signups if costs or abuse spike.
• A status note ready to post if something breaks: what happened, what you're doing.

A prompt for the one AI-appropriate part, rehearsal questions (untested template; use it only for your private prep, not for text you post):

I am about to launch [one-line description] on [Product Hunt / Hacker News].
Here are my fact sheet and the current page text:
[paste]

Act as a skeptical, technically strong visitor. List the 15 questions or
objections I am most likely to get in comments, ordered by how hard they
are to answer. Do not write my answers. Mark any where my page gives no
answer.

Then answer them yourself and keep the notes open on launch day.

Launch day: a checklist, not a strategy

Ahead of time:

• Block the whole day. You need to reply quickly for hours.
• Site, signup, email and billing tested from a fresh browser profile the morning of.
• Monitoring and spend alerts on. Know where to see errors and your API bill.
• Kill switch tested.

During:

• Post at the time you planned. Don't edit-war the title.
• Reply to every comment, including negative ones, with specifics. Thank people for bug reports and fix quickly.
• Don't ask anyone to upvote. Share the link where you'd normally share your work, and ask people to try it and tell you what they think. On HN, don't even ask for comments.
• Keep a running list of bugs, requests and confusing moments.
• Watch costs hourly. Flip the kill switch if the bill moves faster than you can afford.

After:

• Same-day: thank people, fix the top bug, post one short update in the thread if you shipped something.
• Next day: read every comment again and sort into bugs, missing features, and confusions on the page.
• Within a week: write down what you learned and what you will change. That list is the real result of the launch.

What if it flops?

A post can sink with no comments. That happens to good products, and it tells you less than it feels like. Possible causes include timing, title, an unclear pitch, or just a crowded day. Don't delete and repost on HN; the guidelines tell you not to. Instead:

• Ask five people from your target audience to describe the product from your page alone.
• Check whether the demo was reachable without friction.
• Improve it and share it in the places where your specific users already talk.
• Come back with a real update (new feature, new demo) rather than the same pitch. Product Hunt says you can launch again after significant new iterations.

What a launch will not do

It won't fix a product nobody wants. It won't give you durable traffic after the first days. It won't replace a niche audience, an email list or search presence. If your goal is steady users, plan content and distribution that continue after launch day. For the business side of an AI product, the AI wrapper SaaS playbook has the longer view.

Pre-submit checklist

1. I read the current rules on both platforms this week.
2. My project can be tried by a stranger quickly, ideally without signup.
3. I wrote the post and replies myself.
4. I am not planning to ask anyone to vote.
5. Costs are capped and I have a kill switch.
6. Email, billing and analytics are tested.
7. I can be online and answering all day.
8. I wrote down what I will measure and what I will do either way.

Everything above is process. I did not launch anything to write this, so there are no results to report, and I won't invent any.