Self-Promotion
Show off what you're working on
Start new thread

What are you building? Drop your AI product below

by•

Doing one of these because I get more from reading what others are building than from any newsletter.

I'll start.

Building AI Hive, an enterprise AI agent platform that helps mid-market and enterprise teams get from AI pilot to production in weeks instead of quarters. The hard problem we keep solving: compliance, model flexibility, and the lack of in-house AI talent on the customer side.

Product Hunt page:

Your turn. Drop:

- What you're building

- Who it's for

- The hardest part you're solving right now

Will read everything and upvote what resonates.

5.3K views

Add a comment

Replies

Best

Building ErgoVigilance — a computer-vision prototype for ergonomic risk screening.

Who it’s for:

EHS teams, safety professionals, ergonomists, and workplaces that want a more continuous way to review posture-related risk.

It started as pose tracking, but I realised pretty quickly that drawing a skeleton on a screen wasn’t enough.

The harder problem is turning posture and movement into information a human can actually review and act on — risk signals, alerts, timelines, session replay and reports.

Most of the current testing has happened in my room with me as the test subject because I haven’t yet had access to a real factory environment.

I just posted my first proper public demo here:

Still early and not factory-validated yet.

Would genuinely appreciate feedback — especially whether the problem and what I’m building are clear from the video.

Hello!

  • What I'm building: KelpShot, a native screenshot app for macOS. You capture an area, window or screen, mark it up, redact it and share it. It uses Apple's on-device text recognition to copy text out of screenshots and to find words worth redacting (Redact Text covers the ones you pick with solid boxes). Nothing goes to a cloud model, and nothing leaves your Mac.

  • Who it's for: people who take many screenshots for docs, bug reports or support tickets, and don't want another subscription or account. It's a one-time purchase.

  • The hardest part right now: trust without Apple's signature. I haven't paid Apple's $99 signing fee yet, so macOS warns people the first time they open it. That's hard to get past for a privacy-focused app. I publish a SHA-256 checksum for every build and an install guide, and license sales go toward the fee.

Launching on Product Hunt this Sunday:

What I'm building: Holina Factory. You describe a short video clip, it renders on our own GPUs, you keep or toss each take, line up the keepers, and send them to your social accounts. Nothing goes out until someone on your team approves it.

Who it's for: small businesses that need steady short video but don't have a content person. I ran four restaurants before this, so that's who I had in mind.

Hardest part right now: trust. People like the clips, but letting software touch their social accounts is a bigger ask. The approval step helps, but I'm still figuring out what makes someone comfortable connecting their accounts in the first place.

What I'm building: Ovren. You describe a change to your product in plain words, Ovren writes it on a safe copy of your repo, and it comes back as a pull request your team reviews. Next week we add Live Preview: you see the real page with the change on it and click through it before anyone merges.

Who it's for: founders, marketers and PMs with a backlog of small changes stuck behind a dev queue, and the developers who'd rather review a PR than build a banner.

Hardest part right now: trust. Generating the code was the easy half. Convincing a non-developer that the result is right, without making them read a diff, is what we've spent the last few months on. Live Preview is our answer; the launch will tell us if it's enough.

Building DockDuck, a native file manager for Mac: panes side by side, folder rules that tidy your Downloads on their own, iPhone photos that arrive sorted by trip or place, and a page that shows what's filling your disk.

The AI part: if you use the Claude app, you can switch DockDuck on and Claude can work with your files through it. It can find things, rename a batch, sort a folder, and every change shows up in DockDuck with an Undo button. Nothing on your Mac is sent anywhere by DockDuck itself.

Who it's for: everyday Mac users who feel Finder gets in the way, not only power users.

Hardest part right now: Finder is free and already there, so DockDuck has to feel better within the first minute, and stay fast with huge folders and slow network drives.

A task auto-scheduler that's deterministic on purpose. You add tasks with deadlines, durations, priorities - it fits them into your working hours around your calendar. No black box, no "the AI decided" — same inputs, same plan, every time. You can predict and trust every slot it picks.

The scheduling engine has zero AI in it (intentionally), but I built the product solo 100% with AI (Claude Code, get-shit-done toolkit, etc). AI features are coming where they actually help (natural-language task capture, day summaries), but the schedule itself stays predictable.

Honest confession: I love architecting, designing flows, shipping features — and I'm terrible at the part where I tell people about it (try not to think about the product marketing and promotion). This comment is me working on that. 🙂

- free beta has been just opened, I'd appreciate any feedback

What I'm building

ATLAS, a travel app to plan trips, save places you want to explore, and keep track of everywhere you've been. The idea is to bring all your travels, past and future, into one place.

Who it's for

People who love traveling, collect places they want to visit, and have travel ideas scattered across Google Maps lists, notes, screenshots, and conversations. People who want to stay organized without having to plan every detail, and who also want to remember and keep track of the places they've already explored.

The hardest part I'm solving right now

Keeping things simple and resisting the temptation to build every idea that comes to mind! I'm building it solo with the help of AI, and it's surprisingly easy to get carried away.

I haven't launched yet, but I'm documenting the process publicly on Substack, sharing the experiments, decisions, and things I'm learning along the way.

I’m Nafas, the founder of Mivee, a calorie and protein tracker for Android and the web. It’s for people who want food logging, recipes and water tracking in one place.

Manual logging is free, and the free plan includes one AI food scan per day. Scan results are estimates that people can review and adjust. Unlimited scans and planning tools are part of optional Premium.

The hardest part for me right now is discovery: reaching people who actually want to track their food, then helping them understand what they can use for free.

Android:

- What i'm building - Linkedin and X content generation platform that understand you and your brand and generates content that sounds genuinely like you, without the AI slop.

- Who it's for - Individuals and brands trying to build a consistent presence online without the daily block of "what do i post today"

- The hardest part you're solving right now - reaching people who need the product. We are bootstrapped at the moment, so those come with its own limitations.

Product hunt link:

We have a 7-day freemium which you can cancel at any time. Do check it out!


What we've built: - easy self-hosted agents and multi-agent systems.

For: operation managers, BA, COO, automation specialist who want to share agents with team and spent less time for creation, sharing and changes.

The hardest part you're solving right now

Make ai-automation is truly accessible for non-tech users, they don't know enough what exactly can be automated.

First
Previous
•••
293031
Next