Where to launch a developer tool

Most launch advice treats channels as a checklist to be completed on one day. That is the single most common way to waste a launch. Each of these audiences wants something different, each can only get your full attention once, and a flop on one does not have to poison the others if you space them out.

Here is what each channel actually rewards, what it costs, and an order that compounds.

Start by deciding what you have

The right channel depends almost entirely on what a stranger can do with your product today.

Posting a waitlist to Hacker News gets you criticism rather than users, because Show HN specifically asks for something people can try. Posting a rough CLI to Product Hunt gets you silence, because that audience is looking at screenshots. Matching the channel to your maturity is most of the decision.

Hacker News

What it rewards: technical substance, a working demo with no signup, an honest account of limitations, and your presence in the comments. Front page placement is decided in the first hour or two, mostly by whether anyone finds the thing interesting enough to discuss.

What it costs: a day of your full attention, and exposure to criticism from people who have built something similar. It is genuinely one-shot per project per release. You cannot repost the same thing next week.

What you get: the highest concentration of working engineers of any channel, plus a searchable thread that people will find for years. Traffic decays within two or three days.

Who it is wrong for: tools aimed at non-technical buyers, products whose value is the design rather than the mechanism, and anything gated behind a sales call.

We wrote a longer guide to launching on Hacker News covering the title format, timing, and how to handle the thread.

Product Hunt

What it rewards: visual polish and completeness. A gallery of clean screenshots, a short video, a one-line pitch a non-specialist understands, and a maker who answers comments all day. Ranking is per-day, so you are competing against whatever else launched in the same 24 hours.

What it costs: real preparation. Assets, copy, a scheduled slot, and someone available for a full day. It also costs you a favour bank if you lean on your network to show up, which you can only spend once.

What you get: an audience of founders, indie hackers, designers, and product people. Fewer will install a CLI; more will sign up, share it, and write about it. The durable value is usually the backlink, the badge, and being able to say a ranking out loud rather than the raw traffic.

Who it is wrong for: infrastructure with no visual surface, libraries, and anything where the target user is a platform engineer rather than a builder of products. A database driver has nothing to put in the gallery.

Be realistic about the audience mismatch. A tool can top Product Hunt and produce almost no adoption if the people voting are not the people who would run it in production.

Reddit

What it rewards: being a member of the community before you need something from it. Every relevant subreddit has its own rules about self-promotion, some ban it outright, some require flair, some allow it only in a weekly thread. Moderators enforce these inconsistently and unappealably.

What it costs: the most reputation risk of any channel. A post that reads as marketing gets removed, downvoted, or worse, gets a comment thread about your posting behaviour rather than your product. Doing it properly means participating for weeks before you post.

What you get: the most precisely targeted traffic available, and durable search visibility, because Reddit threads rank well and keep sending visitors long after the post drops off the front page. A single well-received post in a small subreddit full of specialists can matter more than a general-audience spike.

How to do it: read the rules, post in the subreddit where your tool solves a problem people are already complaining about, write in first person about the problem rather than the product, and disclose that you built it in the first sentence. Answer questions the way you would in your own issue tracker.

Your own list

What it rewards: nothing. It is the one audience you do not have to win over, because they already asked to hear from you.

What it costs: the months it took to build. This is why it is worth starting a list before you have anything to launch.

What you get: the highest conversion rate of any channel by a wide margin, and the only distribution you own. Platform rules change, algorithms change, moderators change. An email list does not.

Email your list first, a day or two before any public launch. Three reasons: they should not hear about it from a directory, their early feedback catches the embarrassing bug before thousands of strangers see it, and they give your public launch some initial activity that is genuinely organic. If you do not have a list yet, our guide to waitlist page essentials covers the page that builds one.

The channels people forget

These are usually higher value per hour than the big three, because there is less competition.

• Package registries. For a library, npm, PyPI, crates.io, or Docker Hub is where evaluation actually happens. Your registry README is a launch page. People land there from search having never seen your website.
• Directories that developers browse for new tools. A listing keeps working after launch day, which none of the feed-based channels do. Submit your tool with a teaser, a status, and a launch date, and it appears in Explore and on the launch calendar. Listings for tools in beta also land in the try in beta collection, which is where people go specifically looking for something to test.
• Niche newsletters. Many have small lists and unusually high engagement, and most take submissions. A mention in a newsletter written for the exact role you are targeting is worth more than a much larger general-audience spike.
• Adjacent issue trackers and discussions. If people have opened issues asking for the thing you built, a single honest comment linking to your tool is on-topic and welcome. This is the highest-intent traffic you will ever get. Do not spam it.
• Communities you already belong to. Discord servers, Slack groups, local meetups, conference lightning talks. Low volume, high trust.
• Awesome lists and comparison pages. Slow, durable, and worth a pull request.

A sequence that compounds

The order matters more than the set. One workable sequence, assuming a product someone can install:

1. Weeks before. List the tool as upcoming so it starts accumulating search visibility while you build, and start collecting emails. Plenty of developers deliberately follow pre-launch listings and roundups to find tools before they ship, which is free distribution you get for filling in a form early.
2. Two days before. Email your list. Fix whatever they break.
3. Launch day. Hacker News, in the morning you can commit to, with your full attention on the thread. Nothing else that day. You cannot be present in two comment sections at once.
4. Later that week. Package registry README updated, issue-tracker comments where relevant, submissions to newsletters.
5. A week or more later. Product Hunt, once you have fixed what HN found and can show screenshots of something that works. A separate day means a separate audience and a second wave rather than one merged spike.
6. Over the following weeks. One subreddit at a time, spaced out, in communities where you are already a participant.
7. Ongoing. Move your listing to launched, put whatever third-party signals you have earned on your own site, and keep writing about the problem rather than the product.

The reason to spread this over a month rather than a day is not politeness. It is that each channel teaches you something about your positioning, and you want to apply what HN told you before you spend your Product Hunt slot.

What not to do

• Do not launch everywhere simultaneously. You will be absent from every thread and your messaging will not improve between them.
• Do not ask for votes anywhere. It violates the rules on most platforms, it is detectable, and the penalty is silent.
• Do not launch without a way to keep the traffic. If there is nothing to subscribe to, star, or install, the spike is entertainment.
• Do not relaunch the same thing repeatedly. Relaunch when there is genuinely something new: a major version, a new platform, open sourcing something that was closed.
• Do not skip the boring channels. Documentation, a registry README, and a comparison page will outlast every feed post you make.

The honest takeaway

No single channel decides whether a developer tool works. Hacker News gives you a verdict from engineers, Product Hunt gives you reach among builders, Reddit gives you precision, and your own list gives you the only audience you control. The mistake that costs the most is treating launch day as one event to be maximised, rather than a sequence where each step makes the next one better. Pick the one channel whose audience most closely matches your first user, go deep on it, and let the rest follow over the next month.