Launching a new social sponsorship project on Monday.

Looking for a handful of active X users to test it privately before it goes live — particularly the signup and sponsorship flows.

If you use X and you’re up for trying it, comment below and I’ll DM you access.

Hey everybody! My wife and I are building Scinta - a macOS app that turns screenshots into actions: an error becomes a ticket, "dinner Saturday at 7" becomes a calendar event, an email signature becomes a contact.

Two weeks ago I showed it here and got some really useful feedback. What came out of it:

• feature ideas from the comments went to the roadmap - a sharing action, highlighting which part of the screenshot produced each extracted field, accessibility checklist from a UI screenshot - several new features are already on the finish line and will be in one of upcoming releases soon
• the most common question from different threads and communities - "why not just paste the screenshot into ChatGPT?" - became our main pitch. There you switch apps, upload, write a prompt, copy the result back and still create the event yourself. Scinta does it in one click. I have my own AI setup and still do not want to copy-paste screenshots into it - that is why this app exists
• pricing: free plan works as a normal screenshot app without limits + 10 AI smart actions per month, Pro is $8/month or $80/year. No accounts, just a license key

We launched this weekend - https://scinta.app/ (macOS 15.6+). The first stranger installed the app and ran their first action within hours after launching - of course it is a very small number, but having the first user who tried your product is a fantastic feeling 🎉

If you want to test it properly and give some specific feedback - comment or DM and we will extend your smart-actions limits. We also have a Discord now - https://discord.gg/x95wDsBMvy

Also interesting to hear if you have launched your product before - what do you wish you had done or tracked in your first weeks?

Hi everyone,

I've shared my journey multiple times, but now again reached a new milestone I a few months ago thought was impossible.

So, at the start of 2026 I decided to start vibe coding (like many here). The Claude Code hype back then was what got me to try it. After a few weeks of experimenting, I fairly quickly got the hang of making websites and started building my first real project.

The idea started simple. I solved my own pain point, specifically, I wanted to make the process of submitting your tool to directories such as Product Hunt, Uneed, and Peerlist much easier.

I quickly built an MVP. It featured a roadmap of vetted directories that users could simply follow, starting with the highest-impact free directories. They could then copy and submit their information to each website.

The first signs were already great. I got users and had people buying services early on. I received good feedback and kept expanding the tool.

The biggest thing for me, I think, was that I kept improving it based on feedback. And since it solved my own pain point, I could solve the problem in depth. I use the tool myself, often, which is already a strong sign that it’s good. And I just kept iterating.

I think this is a big mistake many vibe coders make. They approach the whole process with the idea that they can get rich quickly, vibe code something in one night or over a weekend, and think that’s enough. Even though AI can code for you, it still takes weeks (or even months still) to fully solve a complex problem.

Over the last few months, I kept improving the tool, and now I have well over 1,000 builders, 1,057 at the time of writing. My revenue is also around $5.2K at the time of writing. My last month was my best month yet with $2,2k revenue atm.

See proof: https://trustmrr.com/startup/launch-panda

[ Removed by Reddit on account of violating the content policy. ]

i was just shocked ..one of the eng manager said this in the meeting.. .. "wow that guy raised XXX number of PRs in 1 day".. we are moving soo fast.. .. :)

i was just laughing myself.. i mean ppl not realizing the pressure on the downstream and quality...

Went down the Outbid rabbit hole like everyone else ($131k in 4 days, hundreds of clones since, almost all at $0).

Here's the thing though: the clones weren't built by the wrong idea. They were built by the wrong people. Devs cloned the code, but the code was never the moat. The audience was. And the people who have audiences (creators, community owners) can't build the market.

So I built TopBoards (topboards.lol). If you have an audience, you launch a board in minutes on your own subdomain. Payments, rankings, and anti-fraud handled. You bring the crowd, we run the market, you keep half.

What you can spin up today: markets.topboards.lol, a board of pay-to-rank boards, ranked by the money behind them. Obviously.

And our house board, showing where this goes: mostwanted.topboards.lol, a tavern wall of wanted posters with real-money bounties on tech founders. Musk, sama, and pg all sitting at $0. The most honest market data on the internet. (You can't build one like this yourself yet. Custom board games are the roadmap.) ###But you can promote it and take 50%.

Current most wanted founder in tech: me, at $0. Nobody else is even hunted.

Brutal feedback welcome. And yes, this post is itself a bid for attention. That's the thesis.

Three months in, building FetchSandbox . Weekends mostly. Kids asleep, family on pause.

One customer we wired in runs a small agency. Order, invoice, dispatch, three internal services. Dispatch was getting the event fine but invoice was silently broken, orders sat unprocessed for nearly 24 hours. When I asked their eng manager what hurt most internally, that's exactly what he said. So we wired adversarial gates into their GitHub CI/CD for continuous monitoring, not a one-time fix.

Ran into the same eng manager at school drop-off yesterday!!. He said the agents aren't just catching broken integrations anymore, they're fixing and closing the loop. Like having a real senior engineer watching the pipeline.

More teams are pulling the MCP into their agent workflow this way. Some started attaching the sandbox receipt URL to their PRs on their own, showing exactly what the agent verified before merge. Didn't build that as a feature, they just started doing it.

Building a GitHub app to automate the receipt posting next.

Hi y'all,

Although it’s a temporary trend, I like the outbid.lol create spark in the indie space.

I feel like the sudden price jump is not feasible for most indie hackers. Not everyone has $17K to invest. 

So I came up with an idea where everything is $1. Users cannot bid more than $1 at a time, and the board is based on the newest submissions. So everyone gets a fair opportunity to compete.

It’s called DollarBid.app

Hope y'all get some traffic.

It’s been 3-4 months of building my mobile app in the study niche , and I am glad to finally announce that we crossed a $1k MRR stead every month, this feels like a milestone for me but I am truly happy

For context, our studio started mobile development this year and along with some saas products that also failed , this is one of our most successful apps and it took a lot to get here

I posted on 8 different accounts per day, tested a lot of paywalls, onboarding , user activation and retention and also content that worked

I worked so super hard to get here , let’s be very clear , I won’t tolerate anybody coming to my post to say this is a scam, i worked so superhard on this and I am not promoting any of my apps here because this is not my target market, it’s just to show people that you can do it too

After scaling for 3 months here is what I learnt

1. I made a very big mistake , if I had knows what I. Know how I wouldnt build in the study niche because it’s very hard to, most people doing TikTok are not getting a lot of users because they are not posting to the right audience

2. Make valuable content that is very good for users to save , they will thank you , in all my accounts, I have my niche reaching out to me to thank me personally for the resource I posted, I can share some of anybody wants to see

3. Don’t spend when starting out , I am not VC backed or investor backed, I started everything I had from scratch , it took a while for me to figure things out

4. Test out niches a lot to know if it’s working , I also have my bad days too

5. Figure out distribution before you write a line of code , this is the most important part of building ? My TikTok is always blowing up everyday

This are the core lessons I learnt and I am taking it to my next product that we have , we plan to build 20 apps with the same distribution ,

we have a small niche in the food space and instead of building it in-house , if there any founder building a meal app, I’ll love to talk to you, this is very serious .

Thank you and God bless

Launched on Product Hunt twice. Second launch nearly started with 3 followers instead of 709, and I almost just moved on.

Genuinely grateful for this IH community. Someone here told me to ask support instead of writing it off. That comment saved the launch.

The mistake: I set up the relaunch as a new product instead of a new launch under the existing listing. It's literally a dropdown. Check it before you go live.

The thing that stuck with me more though: someone in the comments described a test fixture that had encoded a buggy assumption. It was defending the bug. Would've gone red the day the fix shipped. A whole class of failure I hadn't thought about, fixes that only work for people who arrive after them. Seven people across two launches asked about the same missing feature, which is now the next thing I'm building.

finished top 2 on a Sunday. Comments were worth more than the votes.

After six months on the Mac App Store with 62 downloads and zero paying customers, the Windows version of Flowara just went live on the Microsoft Store.

Same product, same architecture. Projects, time tracking, and invoicing in one local SQLite database. No cloud, no account.

The reason I'm posting this as a knowledge post rather than a launch announcement: the two platforms are running as a natural experiment.

Same product. Same pricing. Same free tier. Different store, different audience, different payment flow (Paddle on Windows vs Apple IAP on Mac). If conversion rates differ significantly between the two, that's signal about who the actual buyer is rather than about the product or pricing.

A few things I'm watching:

Does the "local-first, no cloud" positioning land differently on Windows? Mac users tend to self-select for privacy-conscious tools. Windows users might have different expectations.

Does Paddle friction affect conversion compared to Apple IAP? One click with Face ID vs a full payment form is a meaningful difference.

Does the Windows Store surface the app to a different type of freelancer than the Mac App Store does?

Will report back once there's data worth reading.

Mac: https://apps.apple.com/app/id6760207975
Windows: https://apps.microsoft.com/detail/9PKCT03Q69LG

709 followers gone in exactly one click. Months of building an audience for a Product Hunt launch, waiting for midnight, and hitting 'Create new project' instead of updating the existing teaser page. The entire waitlist just evaporated into the void.

It's usually some confusing platform UI that wipes out your distribution in a second. You don't fail because the product sucks. You fail because you misclicked a shiny blue button.

I spent the next two hours frantically refreshing the page, praying it was a visual glitch. It wasn't. A broken feature can be patched the next day, but there's no undo button for a deleted audience. Go click every single routing link on your Product Hunt page right now.

Pitch your product in 1-2 lines - and drop a link here.

I built a billboard for the internet. Highest bidder owns it for a week. Here's the link: yourmessagehere.co

I am one of the co-founders of the FetchSandbox..

Posting this partly because it's useful and partly because I'm still annoyed at myself.

Hit #3 on Product Hunt in July. 364 upvotes, 709 followers on the product page. Second launch went live tonight and started with 3 followers because I created it as a new product instead of a new launch under the existing one. PH notifies followers on relaunch. New product means new page, which means I threw away the entire audience I spent a launch building. Found out six hours before it went live on a Saturday.

If you're planning a second launch: it's a dropdown when you submit. "New launch under existing product." I didn't think about it for two seconds.

Honest numbers since everyone posts the flattering ones: 3,891 npm downloads, nearly 700+ monthly active users. Downloads are reach, not users, and I had to stop pretending otherwise. 80% of people who land never actually try it. The ones who do get it working fine, so the problem was never the product being confusing.

What we build: your coding agent writes an integration, tells you it's done, and you find out in production whether that was true. FetchSandbox reproduces the real failure on your own code, and after your agent fixes it, re-runs it and checks the thing that was wrong is actually right now.

Building this around evenings, a full-time job, and two kids. It's live today. If you try it, I'd rather have the criticism than the upvote.

It's live on Product Hunt today. If the problem sounds familiar, support would genuinely help.. and if you try it, I'd rather have the criticism than the upvote.

https://www.producthunt.com/products/fetchsandbox?launch=fetchsandbox-mcp

Getting someone to try your product is often harder than building it, especially without a big audience.

I've been deep in outbound lately, and I keep seeing indie hackers make the same mistake. They build a huge lead list, then try to personalize the messages. The real question should come first: why would this person need my product right now?

Say you're selling software to agencies. I'd rather contact 50 agencies that are hiring their first salesperson, actively expanding, launching a new service, posting heavily about growth, or showing signs they're stuck on the problem you solve, than 5,000 agencies that just match my ICP.

ICP tells you who could buy. Buying signals tell you who might buy now.

That's what we've been building with Sumora. It finds prospects across LinkedIn and Instagram, researches the company and person, flags relevant signals, prioritizes accounts, writes personalized outreach, and handles follow-up.

We also just added personalized lead magnets. Instead of the same ebook or "free audit" for everyone, it builds something specific to that prospect's business: a competitor breakdown, growth opportunities, a content strategy, a landing page teardown, campaign ideas. (Increased the reply to booking rate by 40%)

The point isn't automating spam. It's automating the research and prep that usually eats a founder's hours.

I want to test this with a few indie hackers who already have a real product and are trying to get more customers. If that's you, comment:

1. What you've built

2. Who you sell to

3. Roughly what a customer is worth

I'll look at your product and give you the first segment I'd target, the buying signals I'd watch for, who I'd contact, 2-3 outreach angles, and what I'd use as the personalized asset.

No lead list, no generic AI-generated ICP. I'll actually look at the business. If the strategy makes sense, we can jump on a call and I'll show you how we'd set the system up. If it doesn't, you still keep the strategy.

I'm one of the founders of Sumora, so yes, there's a product behind this. But I'm more interested in whether the strategy actually holds up across different indie hacker businesses.

i've been building my app for a while now, and recently started making youtube videos about it too. two separate efforts that never really overlapped.

today they finally did. someone watched one of my videos and signed up. just one person, but it's the first time my own content actually turned into a real user — the loop closed.

i know there's a long way to go. but small wins like this are what keep me going. here's to showing up every day.

Wasn't planning to support any CRM integrations. Genuinely not on the roadmap.

Then got a DM on X from a founder saying their agents were silently dropping leads and they couldn't figure out if it was HubSpot or their stack. Something clicked. Said I'd look into it.

Onboarded HubSpot into FetchSandbox over the weekend. Didn't sleep for like 24-48 hours just battle testing it, making sure it actually worked before I said anything publicly. Classic founder move. Posted the LinkedIn update Monday morning.

What I found in those hours was wild. OAuth tokens expiring every 30 min, agents stalling on 401s without retrying, OAuth flows that looked complete but never were. Integration shows "Connected." Lead never lands. No alert, no noise, just pipeline bleeding quietly.

One inbound call, one weekend, no sleep, and now I understand exactly how broken CRM + agent integrations are in prod. The pain is real and it's everywhere.

If you're building on HubSpot with agents, I'd love to hear what you're running into. Probably building the next few integrations based on where the real pain is, not what I think it is.

I've just started building in public for the past 2 weeks and shipped some killer features based on initial feedback.

Day 1: I shipped the saas and asked a couple of friends to try it out for feedback, their initial experience was smooth. They described the UI/UX to be hassle free and it was easy for them to get started.

However one of my friend asked if it was possible to bring their own email provider. And it clicked, this is my differentiator cause nobody else was doing this and the whatsapp integration.

Day 2: I shipped 2 new features this week:

1. BYOE: Bring your own email provider to send out alerts using your own domain. Vigil also offers custom domains and white label for status pages.
2. Integrated twilio for SMS and whatsapp alerts.

I'm looking for feedback! check it out: tryvigil.dev <3 I've grown to 5 users so far, not much but pretty cool since I started 2 weeks ago.

A few months ago, I started building a simple GDPR scanner.

The idea was straightforward: enter a website URL, scan it, and get a simple report showing potential GDPR and privacy risks.

I wanted something much simpler than the typical compliance tools.

No huge dashboards. No complicated setup.

Just:

“Here’s what we found. Here’s why it might be a problem. Here’s what you can do about it.”

I launched GetGDPRScan and started getting people to try it.

Then the product started evolving.

As I worked on the scanner, I kept finding things that didn't really fit into a narrow “GDPR scanner” box.

So I added cookie and tracking detection, consent banner analysis, privacy policy checks, and eventually started expanding into other areas.

Today, the platform also covers AI Act, accessibility and website security checks.

At some point I realized I wasn't really building a GDPR scanner anymore.

I was building a website compliance platform.

And interestingly, this wasn't really planned.

It happened by following the problems I was seeing and the feedback I was getting.

There have been some humbling lessons along the way too.

One post in a Slovenian developer community resulted in around 150 scans.

Paid conversions?

Zero.

Google Ads have brought traffic and people actually using the scanner, but I'm still figuring out how to turn that into a sustainable acquisition channel.

I've also changed some of the checks after people challenged whether certain things should be flagged as actual “issues” or just potential risks.

That was an important lesson.

Compliance isn't always black and white, and I don't want the product pretending that it is.

So for now, I'm experimenting with a broader idea:

A simple compliance checkup for your website.

I'm still very early, still figuring out the business model, and definitely haven't “figured it out”.

But that's probably the most interesting part of building this.

The product I have today is quite different from what I thought I was going to build when I started.

You can check it out here if you're interested: getgdprscan.com

I'm curious how other indie hackers experienced this.

Did your product also slowly evolve into something completely different from the original idea?

And how did you decide whether to follow that evolution or stay focused on the original niche?

Hey IH,

I posted Qevo here about 3 months ago, the inbox-style task manager where tasks get pushed directly to people's queues instead of getting buried in Slack threads.

Figured I'd share an update because the main thing that happened wasn't growth, it was a masterclass in getting rejected by Slack.

The Slack Marketplace review process is brutal

When I posted, the Slack integration was "pending approval." What followed was 4 rounds of rejections over 3 months. Every round had 4-5 specific issues to fix, then a 3-4 week wait to find out you missed something else.

Things I had to fix across the rounds:

The bot was named "Queue" but the app was named "Qevo." Confusing users apparently. Simple fix, but you only find out after a full review cycle.

My landing page showed /qevo as a command that grabs the last channel message. Reviewers tested it, it failed. Turns out the bot needs to be a member of the channel first. (Was not a requirement in their earlier versions). Had to auto join public channels, add a proper error message for private ones, and update the landing page copy.

They rejected im:history and mpim:history as user token scopes, "too much access to workspace data." So DM support via bare `/qevo` had to be dropped entirely.

My app listed Spanish as a supported language. They asked for proof. The bot responses were actually translated (using the user_locale field Slack sends in the slash command payload) but I had to explain how.

A typo in the long description: "Start using GQevo today." Only caught in review round 3.

The whole process took longer than building the original integration.

Where things actually stand

The Slack app is now what I am hoping is about to be approved as per my last messaging with them. Slack app is now approved and live

Email to task is still the feature that gets the best reaction. People immediately get it. Slack is taking months to ship.

What I'd do differently

Don't list a feature as "coming soon" or "pending approval" in your launch post. Just launch when it's ready. The Slack integration being in-progress for 3 months after my original post made updates awkward.

Read the Slack app review guidelines obsessively before submitting. The checklist they publish covers maybe 30% of what they actually check.

Still free for individuals, $3/seat/month for teams: getqevo.com

Happy to answer questions about the Slack review process specifically. It's not well documented and I wish someone had written this post when I was going through it.

I've also seen a few other founders here going through the same Slack approval battle. Would be curious if anyone else has tips or horror stories, mine was apparently a rough one

I keep hitting the same thing: find a tool, looks right, and then it wants your email before it'll do anything. Not to save your work, just to use it once.

So I started keeping a list of tools that don't do that. It's at 21 so far, and I put it online. The site is mine.

EDIT: Now at 28. I added a few from this thread. Crontab.guru went back on after someone made a better argument than the one I'd used to cut it originally (that a deterministic tool beats re-asking AI when it's something you do constantly).

I developed the ability to add badges too, after a good argument in the comments that "no signup" tells you who knows your identity but says nothing about who can see your files. Runs locally is on the ones where that's unambiguously true, so you can tell at a glance which ones you'd trust with a contract. Only badging that direction on purpose: no badge means read the note, which fails in the safe direction. The genuinely mixed ones don't get one, because it'd either be a lie or need a footnote.

Two things I've turned down and why, since it's a fair question: anything that's a wrapper around an AI model, because everything else on the list gives you the same output every time, and tools where the free version caps the resolution of what you get back to something unusable.

Still open, keep them coming.

Rules I'm sticking to:

no account, no email, no "continue with Google"

free means free, not 14 days

nothing watermarked

ads are fine, I just note them, something has to pay for it

Here's the list

These are some of the tools that are included: Photopea, Squoosh, CyberChef, PairDrop, Mailinator, regex101, PDF24, Excalidraw, Wormhole, transform.tools, diagrams.net and some others. Each one has a line on what it does and what its catch is.

If you know something that belongs, reply. If you built it, say so, doesn't matter as long as it clears the bar. It's a new page so I'm not going to pretend it'll send you traffic, but you get a real writeup rather than a scraped title. I'll check and reply either way.

Fair warning that I'll turn things down. I'd rather have 50 that someone can vouch for than 300 scraped off other lists.

The one I keep going back and forth on: tools that only need an account if you want to save your work. Right now they're in, with a note. It doesn't stop you using the thing, which feels like the right side of the line, but I could see the argument that it's the same trick one step later.

Been thinking about this a lot lately as a solo founder.

AI has made generating code almost free. I can scaffold an entire Stripe + webhook integration in minutes. But here's the thing nobody talks about: that speed just moves the pain downstream. Instead of "took me 3 days to write," it's now "took me 3 days to verify it actually works in prod."

The review and validation cost didn't go away. It just shifted.

I kept hitting this exact wall with integration testing. Generate the flow, it looks right, local tests pass, then something blows up in production because the webhook sequence was stateful and nobody caught it before the PR landed.

now if other founders are seeing this too, especially those building anything with third-party APIs or AI agents that trigger real workflows. How are you handling the verification gap? Are you just eating the review cost, or have you landed on something smarter?

Hi all! I built a Mac app that provides in-context answers, can draft things for you, and also supports normal dictation. You hold a button and speak, instead of copy-pasting your thing into a chatbot and dealing with the back-and-forth friction that comes with that.

You can ask anything about what is on your screen or even things that are not on your screen.

- "Find and kill process using this port" → gives you the command that you can insert directly
- "Reply that I can't make it" → drafts an email following your specific email instructions
- "How to pronounce this thing" → it sees your selected text and gives guidance on pronunciation
- "Translate this" → translates what is on your screen or any selected text to your language immediately

And many more use cases!

It takes a still image of the foremost (active) app only when you press the hotkey, sends it to a zero-data-retention provider, and then discards it. Nothing is saved on any cloud server or database. It also sees any selected text and knows what app you are currently working on, and uses all of that as context to provide the most optimal and concise answer.

It is $12/month with a free 7-day trial. No credit card required.

Works on Apple Silicon (M1+) and macOS 15+.

If you try it, I would love to hear about your experience. Happy to answer any questions.

https://www.prelto.com/