This is the playbook. It assumes zero followers, zero email subscribers, and zero press contacts, because that is the default starting position.

One working example I will use: StartupWiki, built by Pranesh, runs a weekly launch board for newly shipped products, ranked by community upvotes with a decaying boost for recent launches. It sits alongside a public research directory of startup profiles, but the launch board is the part that matters on day one.

Why does launching with no audience feel impossible?

Because the launch advice most founders read was written by people who already had distribution. "Announce it on Twitter" works when you have 40k followers. "Send it to your list" works when you have a list. When you have neither, launch day feels like shouting into an empty room.

The reframe that fixes this: an audience is not a prerequisite for launching. It is one of several distribution channels, and it happens to be the only one you build slowly before launch. Every other channel can be used on launch day with no prep:

1. Communities where your users already gather.
2. Launch boards built for discovery.
3. Directories people search when they have a problem.
4. Direct outreach to the 20 people who have the problem worst.

Where do you find users when you have no audience?

1. Niche communities. Subreddits, Discord servers, Slack groups, and forums where your target user complains about the problem you solve. Do not drop a link on day one. Answer questions for a week, then share what you built when it is genuinely the answer.
2. Launch boards. These exist specifically for makers with no audience, because the traffic comes from the board, not from you. More on these below.
3. Directories. Slower, but they compound. A listing keeps working while you sleep.
4. One-to-one outreach. Find 20 people who have the problem badly and write to them directly. With no audience, your first users are recruited, not attracted.

How do launch boards work when you have zero followers?

A launch board flips the normal equation. Instead of you bringing traffic to a listing, the listing sits in front of traffic that is already there, because people browse launch boards specifically to find new products.

Using StartupWiki as the example, the flow is: create a free account, upvote one of two projects the board shows you, submit your product, and add the "Featured on StartupWiki" badge to your own site. Once the badge is verified on your site, the launch goes live. There is no editor approving or rejecting you; verification is mechanical.

Ranking matters more than posting. Launches are ranked by community upvotes with a decaying boost for recent launches, so a brand-new product gets real visibility while it is fresh, and that boost fades rather than expiring overnight. Your launch gets its own page and stays on the board, so it keeps working after launch week instead of vanishing in a 24-hour news cycle.

What should you do in the first 48 hours?

1. Post to your chosen communities in the morning, when your target users are active.
2. Submit to a launch board and make sure the badge on your site verifies, so the listing goes live without a delay.
3. Message your 20 handpicked prospects personally, one by one, with a note about their problem rather than your product.
4. Answer every comment and question the same day. On a launch board, maker activity is part of the signal browsers use to judge a product.
5. Write down where every signup came from. With no audience, you cannot afford to guess which channel worked.

What mistakes kill a no-audience launch?

• Launching only on your own channels. A tweet to zero followers is not a launch.
• Skipping the community warm-up. Dropping a link into a community you never contributed to reads as spam, and it is.
• Treating launch day as one day. The founders who win without an audience treat launch week as a full-time distribution job.
• Waiting for a bigger audience before launching. The audience comes from launching, not the other way around.

FAQ

Can you launch a startup with zero followers? Yes. Followers are one distribution channel. Communities, launch boards, directories, and direct outreach all work with zero followers, because the audience belongs to the platform, not to you.

How long does it take to get your first users this way? Founders who work the channels deliberately usually see the first signups within days, not months. The compounding part, SEO and directory traffic, takes longer, which is why you start it on launch day.

Is a launch board worth it for a tiny product? That is exactly what launch boards are for. A board like StartupWiki's is built for small teams with new products, and ranking by upvotes with a recency boost means a tiny product launched this week can outrank a bigger one launched last month.

Should you build an audience before your next launch? Yes, but do it in parallel, not as a gate. Every launch on a public board is also audience-building: the people who find you there are the start of it.

The bottom line

Launching with no audience is the default, not the exception. Borrow distribution from the places your users already are, put your product where people browse for new things, and recruit your first users by hand. If you want a place built for exactly this, the StartupWiki launch board is a good first stop.

What does a failed Product Hunt launch actually look like?

It rarely looks like a flop on the day. A typical failed launch gets a few dozen upvotes, some nice comments, a traffic spike that flatters your analytics, and then silence. Two days later the founder checks the dashboard and daily visitors are back to where they were before launch.

The painful part is that nothing went "wrong". The product did not get worse on day two. The platform simply stopped showing it to anyone. That is the failure mode worth understanding, because it is fixable.

Why does visibility decay so fast on launch platforms?

Recency-ranked feeds are the reason. Product Hunt, and most launch platforms like it, order products by when they launched, with votes mostly deciding placement within that day. Once your day ends, no amount of upvotes puts you back on the front page. The algorithm is doing exactly what it was designed to do: feature what is new, not what is good.

Social posts decay the same way. Your launch tweet has a half-life of hours. Your friend's LinkedIn post about you lasts a day. Every channel you used on launch morning is a channel with an expiry date.

What are the most common launch-day mistakes?

1. Treating launch day as the finish line. The launch is planned for weeks, and nothing is planned for the day after. The spike arrives with no system to catch it.
2. Relying on friends for upvotes. Friends vote once and disappear. Their upvotes buy a few hours of placement, not a distribution channel.
3. No follow-up assets. No "we launched" recap, no numbers post, no answer to the question everyone asked in the comments.
4. Ignoring the comment section. Every unanswered comment is a warm lead gone cold. Launch-day commenters are the most convertible audience a founder ever gets.
5. Assuming one platform is the whole plan. Product Hunt is one spike in one community. Founders who stop there have bet their entire launch on a single 24-hour window.
6. Expecting a failed launch to mean a failed product. A weak launch usually means weak distribution, not a weak product.

What separates launches that keep growing after day one?

Look at products that are still getting users weeks after launch and the pattern is consistent: they converted a temporary spike into permanent placements. They rank in search for a problem their product solves. They are listed in directories and launch boards that do not expire. They captured emails on launch day and turned one day of attention into an audience they own.

None of that happens by accident on launch morning. It is the day-after work, and it matters more than the launch itself.

How do you keep a launch alive after the spike?

The highest-leverage move is placing your product somewhere designed for permanence rather than recency. StartupWiki is a weekly launch board for newly shipped products, ranked by community upvotes with a decaying boost for recent launches, plus a public research directory of startup profiles covering funding history, financial metrics, competitors, leadership and dated sources. Pranesh builds StartupWiki to fix exactly the failure mode above: when you submit a launch, you verify a "Featured on StartupWiki" badge on your own site, your product gets its own page, and it stays on the board with no 7-day expiry. New launches still get a boost, but your upvotes keep compounding instead of resetting at midnight.

If your last launch disappeared after 48 hours, close the gap now: add it to the StartupWiki launch board, then build the rest of your always-on channels on top.

FAQ

Can a Product Hunt launch succeed after a weak first day?

Yes, if you stop measuring the launch by the day and start measuring it by what the day built. A weak spike that leaves you with emails, search rankings and permanent listings beats a strong spike that leaves nothing.

Is it too late to list a product that launched weeks ago?

No. Boards and directories that rank by community votes rather than pure recency are built for exactly this case. A product listed late can still climb and, more importantly, it stays findable.

Should I relaunch on Product Hunt if the first launch failed?

Only when you have a genuinely new story to tell, such as a major feature or milestone. Relaunching the same product because the first spike faded burns community goodwill and does nothing to fix the underlying discovery problem.

Product Hunt launches fail because founders plan for a day and the internet rewards what lasts. Plan for both.

Honestly though, Openclaw was super overhyped. After trying it, it took me forever to get it set up, and even after that it broke often. Connecting Gmail, calendar, etc. felt impossible, and to be honest, Openclaw lost me more time than it saved.

When I saw Muse being released, I was confident it was going to be the same, but after a week of using it, it genuinely stood out to me.

The app has gone super viral, hitting number 1 on the app store.

In this article we will cover:

• The basics of how Muse works in the backend (skip if you wanna get to the usage part)
• The Muse UX and its parts
• How to set Muse up
• Best ways to use it (don't miss this part!!)

How Muse works

Muse's basic concepts are very similar to AI agents, but there are some key things in this architecture that made it go so viral and be so useful.

The first part is the Agent itself—this is just an LLM being called with an API request + context. Every time you talk with Muse, it adds important details to its Soul.md (which defines personality) and its Memory.md (which records important info).

The second thing is by far the coolest, and the fact that Meta gives it away for free is even more insane. Muse can actually use a computer, each Muse agent gets its own virtual computer along with a stateless browser.

Here's how this works: if you tell ChatGPT to help you get a job at an AI startup, it's going to search jobs in AI startups and then summarize the results. If you tell Muse the same thing, it's going to go onto its virtual computer, log in to your LinkedIn to understand your career trajectory better, then look at your connections to figure out who you know best who can warm intro you into a job. Then it can tell you why those people are good fits, give you a template for what to message, and rank the outreaches by chance of getting a job. Same prompt, but a very different result.

The next big feature with Muse is safety and ease. For other AI agents, you have to create and link APIs if you want to access things like Google Calendar. With Muse, you just click connect and an auth connects you.

The best part about Muse is that, unlike Openclaw or many of these tiny startup AI agents, Muse actually is secure. It stores all of your logins in a secure folder that the AI agent never sees, so you can guarantee that your info is safe.

The Muse UI explained

Muse's interface is clean and minimal, designed around a central chat window where you communicate with your agent. Here's what you'll see:

Chat Interface - Pretty standard, this is where you talk with your agent

Sidebar Navigation - On the left side, you'll find quick access to different sections: Home (main chat), Tasks, which are ongoing automations, Memory (what your agent knows about you), and Settings (app and agent configuration).

Agent Status Indicator - A small icon showing whether your agent is idle, actively working, or waiting for your input. When it's working, you can see a loading state and sometimes a brief description of what it's doing.

Connected Apps Section - Shows all the apps you've authenticated (Gmail, Google Calendar, LinkedIn, etc.) with small icons and a toggle to enable/disable each one. You can add new connections with a simple "Connect App" button.

Memory & Soul Panels - Tucked into the settings, these let you view and edit your agent's personality (Soul.md) and what it knows about you (Memory.md). Both are displayed as readable text files you can scan or update.

Task Cards - When your agent is running recurring tasks (like a daily digest or weekly outreach), they appear as cards showing the task name, next run time, and status.

Feed - This is a place your agent communicates with you without needing to clutter the main chat, for example if your agent needs to remind you it will put it in the feed.

Overall, the UI is not that hard to navigate, I recommend just trying the app and you will have it down in a few days.

That covers how Muse works and what the app looks like. If you want to go further, I put the setup walkthrough and the exact copy-paste prompts (the interview prompt, the LinkedIn outreach automation, and the daily digest) in a bonus guide on my Substack: The ultimate guide to using Muse (with prompts).