How to launch a SaaS on Hacker News

Hacker News is not a launch button for SaaS. It is a community that rewards curiosity and punishes marketing copy. If you treat Show HN like Product Hunt day with a friend army, you will get flagged, buried, or roasted. If you show something people can actually try, write like a peer, and stay in the thread, you might get useful, blunt feedback.

In 2026 anyone can ship a weekend product. The feed is full of lookalike tools. Pieter Levels put it plainly: building got easy, attention did not. HN will not fix a product nobody can use, and it will not forgive a pitch that reads like a homepage. Indiecity is for one person or a team under twenty. This guide is how you post without spamming the room.

What Show HN is for

Read the official page before you submit. Show HN is for something you made that other people can play with. Users try it, give feedback, and ask questions in the thread. On topic: things people can run on their computers or hold in their hands. Hardware can use a video or a detailed article. Books can use a sample chapter.

Off topic for Show HN: blog posts, sign-up pages, newsletters, lists, and other reading material. Those cannot be tried, so they cannot be Show HNs. Make a regular submission instead if the material belongs on HN at all. Do not post a one-off you generated this morning. Anybody can do that now. Share something you actually worked on, and explain how and why.

You must have worked on it personally, and you must be around to discuss it. It does not need to look slick. Early work is fine. New features and minor upgrades ("Foo 1.3.1 is out") generally are not enough. A major overhaul might be. Landing pages and fundraisers are out.

Is the product ready?

If people cannot try it, do not do a Show HN yet. Come back when they can. The same page asks you to make it easy to try, ideally without signups or email gates. You get more feedback that way. dang’s long tip comment says the same: the project must exist, people must be able to try it, and signup walls annoy HN readers.

• A working app, demo, or download people can poke at today
• Clear path from the link to the thing (not a waitlist-only page)
• You free for most of the day to answer comments
• Pricing visible if you charge (opaque pricing draws complaints)
• No surprise sales call gate unless you said so up front

If you are only collecting emails, it is too early. That rule shows up on both the Show HN page and the Launch HN instructions (Launch HN is for YC companies, but the product advice still applies). Hard tech that takes years is the rare exception on Launch HN. Your micro-SaaS is not that exception.

Write like a peer, not a press release

The writing rules are not optional culture tips. They are how posts survive. dang’s tips for presenting work still hold in 2026, with a new hard line: write your text by hand. Do not use an LLM to generate or edit what you post to HN. The community is fussy about LLM writing right now.

Drop any language that sounds like marketing or sales. On HN, that is an instant turnoff. Use factual, direct language. Personal stories and technical details are great.

The Launch HN page spells out the same voice for anyone writing about a product on HN. Do not write marketing, sales, or PR. Talk to readers as peers, not as users or investors. If it sounds like your homepage or pitch deck, it will not work. Tell them what you built. Do not sell them. Use your own voice. Skip fastest, biggest, first, and best. Say what is still rough.

Include a clear statement of what the project does. If you skip that, the thread becomes "I can't tell what this is." Include the backstory of how you came to work on it and what is different. Link any previous relevant HN threads. Put the text in the submission so it sits at the top; if it does not show, add it as the first comment.

1) Who you are in one line.
2) One clear sentence: what it does.
3) The problem: why it is hard and why it matters.
4) How you hit the problem and decided to build this.
5) How it works, with real technical detail.
6) What is different, with real technical detail (not a feature list).
7) How you make money, or what you are thinking.
8) Open ask for feedback, experiences, and ideas.

Do not repeat the title word for word in the first sentence. Find a second way to say what you do. Do not write single-sentence paragraphs that read like a sales letter. Do not stack feature bullets and call that a story. Assume readers are smart. If you use jargon, define it once.

Title and submit

To post, submit a story whose title begins with "Show HN". Every Show HN lands on shownew. After a small points threshold, it can appear on the show page. There is no guaranteed front-page placement for non-YC posts.

Keep the title dry. The site guidelines say not to make titles stand out with uppercase, exclamation points, or praise. Do not editorialize. Keep it modest and factual. dang’s examples lean plain: what it is, not how amazing it is.

Show HN: [Product] - [one factual clause about what it does]
Show HN: [Product], a [short category] for [who / job]
Avoid: "revolutionary", "best", "finally", "game-changer", ALL CAPS, and trailing !

Use a human username, not the company name. Put an email in your profile if you are comfortable so moderators can reach you. dang notes they sometimes send repost invites that way. Read and comment on other people’s posts. Posting your own work some of the time is fine. Turning HN into your ad channel is not.

Do not spam: the rules people break

This is the part most "HN growth hacks" get wrong on purpose. The official rules are not vague.

• Do not ask friends to upvote or comment. Show HN says that is not ok. The FAQ bans asking for upvotes or comments for promotion.
• Do not plant booster comments from friends, users, or coworkers. Launch HN says readers will sniff them out, flame you, and the team may bury the thread.
• Do not delete and repost to reset the clock. Deletion is for posts that should never have gone up.
• Do not solicit votes, comments, or submissions in other channels as a campaign. Vote because you personally find something interesting.
• Do not paste generated or AI-edited text into the post or comments. Site guidelines and dang’s 2026 note both draw that line.
• Do not resubmit the same launch over and over. Significant change only, rare, with a link to the last thread.

Those rules exist because the software and the community both punish rings and fake praise. "Ask five friends from different IPs to upvote in the first hour" is not a tactic. It is how you get penalized. If someone tells you that is the HN playbook, they have not read the FAQ.

Stay in the thread

Reply when comments appear. Take breaks, but do not leave the thread cold for long. Launch HN says that if you disappear, people get annoyed and the post drops faster. Answer with real technical detail. People want the problem, the stack, the tradeoffs, and how you make money.

Do not greet people by username to sound close. Answer the point. When someone criticizes you, find something true in what they said first, even if it is only the intent. You never have to convert a critic. The silent readers watch whether you get defensive. Cheerful, concrete replies win more than a perfect rebuttal.

In comments generally, the site guidelines still apply: be kind, converse curiously, skip snark and shallow dismissals. Show HN’s comment rules ask people to respect the person sharing work, ask questions out of curiosity, and suggest alternatives instead of cross-examining. You set the tone by how you answer the first hard comment.

Launch day checklist

1. Read showhn.html, newsguidelines.html, and dang’s tip thread the same week you post.
2. Confirm a stranger can try the product without a sales call or a pure email gate.
3. Write the body by hand: problem, backstory, how it works, what is different, money, ask for feedback.
4. Dry title that starts with Show HN. Human username. Email in profile optional but useful.
5. Block the day. Answer fast. No booster campaign. No upvote asks in Slack, Twitter, or group chats.
6. Log every concrete bug and objection. Ship fixes while the thread is still alive if you can.
7. When the spike ends, follow up with people who tried it. A Show HN is not your customer list.

After the spike

Upvotes are not customers. A busy thread is not proof people will pay. What you keep is specific feedback and a short list of people who actually used the thing. Write them a human note. Offer a clear price. Ask if they want to keep going. That is the same work as getting your first ten paying customers, just with a few more strangers who already saw the product. When the spike dies, do the week after, not a second post.

If the post went nowhere, do not treat silence as a ban. Most Show HNs get little attention. Fix what was confusing. Make the try path shorter. Come back only when the product is meaningfully different, and say so with a link to the old thread. Between posts, be a normal participant on other people’s stories. If you have no audience at all, launch with named people first. Curiosity first. Promotion second.

What to stop doing

• Show HN for a waitlist, deck, or blog announcement.
• Titles with hype, ALL CAPS, or exclamation points.
• Homepage paste as your post text.
• LLM-written blurbs and comments.
• Friend upvote campaigns and planted praise.
• Delete and repost the same launch.
• Company-name accounts that never comment elsewhere.
• Treating points as proof people will pay.

Then you have a real launch story

A clean Show HN is simple: something people can try, a factual story, a full day of honest replies, and zero spam. The rules are public on showhn.html, the site guidelines, the FAQ, and dang’s tips. Follow those and you belong in the room. Break them and you teach the community to ignore you next time.

If you ran the day, took the heat, and got real users or hard lessons out of it, write it down. When the product is real and the name is yours, put it on the map. You can join Indiecity while you ship the next version. The community is for builders. The customers still come from the product and the ask, not from a single orange link.