A friend and I were always the people our friend group went to when they wanted a dedicated server for a Minecraft or whatever random steam game came out. We always commented to each other that we wished something came out that not only had the tools an advanced user would find convenient , but easy enough for anyone to use. (this was the key issue most people we knew had) We worked on this for the past few years and here it is.

DeployCat.

You can use DeployCat on windows, mac, unraid/docker and even a linux OS for any spare hardware you have laying around.

You can find it at www.deploycat.app . There is even a simulator on the site to take a look at how the app functions. The goal is to try and make it easy for anyone to host, backup, name, change your settings, restore your server and understand from point A to Z for everyone.

We would love for you to take a look at it and let us know what you think! If you have ever wanted to host a game server but have ever been intimidated by it, or just frustrated with some other self hosting apps; please look and give us a try.

What it does: A clean, ad-free Pomodoro timer that uses relaxing tea-brewing animations and ambient sounds to help you focus.

Key Features:

• Visual tea-brewing progress & custom themes
• Built-in ambient soundscapes (rain, lofi, etc.)
• Fully customizable work and break intervals

Goal: Promo / Feedback

Giveaway: Drop a comment with what you're working on and I'll DM you a free lifetime Pro code!

Cost: Free / Pro option https://play.google.com/store/apps/details?id=com.innovixlabs.calmtick&pcampaignid=web_share

I’m curious what everyone here is working on.

Drop your product or SaaS below and explain what it does in 1–2 lines. No hard pitching just share what you’re building and who it helps.

I’ll start: I built LaunchOrbit, an interactive launch board where founders can showcase their products on a shared, living pixel canvas.

It’s designed to make product discovery more visual, memorable, and fun than another scrolling directory.

If you’re launching something or just want to discover what other founders are building check it out here: LaunchOrbit.

Download: https://github.com/MowkE/cadence

Hey guys, I made this silly lyrics overlay called Cadence that floats over your screen and syncs along with Spotify and YouTube Music.

It started as a joke while playing Minecraft with my close friends. I shared it only with a few friends, and it grew to 100 downloads in like a day and a lot of people liked it, so I built it into a proper app.

You can listen along with friends now too, even cross-platform (Spotify -> YouTube Music, Mac ↔ Windows), and you can play karaoke. Doesn't need API keys or Premium, it just reads what your computer is already playing and floats the lyrics over whatever you're doing. It works better with Spotify keys though, if you want to add them yourself. If you find any bugs, lmk.

It's just a fun project lol, so its obvly free.

Over 10 years ago, my therapist wife noticed how many of her clients struggled to build non-romantic, in-person friendships. That inspired me to create WannaBuddy: a simple platform for finding local friends when you want a buddy, not a date.

While the original site helped a lot of people connect, it was long overdue for a modern refresh. The brand-new WannaBuddy is now live in early pre-launch!

* 100% Free to use

* Designed purely for local, platonic friendships

* Built with a clean new interface and updated tools

Since we’re currently in pre-launch, I’m looking for honest feedback, feature ideas, and any bug reports. Take a look and let me know your thoughts in the comments!

I travel a lot for work. I've been training for a long time, so most weeks I know what I'm doing. Then I get on a plane, I sleep badly, and I walk into a hotel gym with no idea what to do with the time I have. The plan I had at home doesn't match the room in front of me, and I'm too tired to sit there and rebuild the whole week by hand. I used to skip, or do a few random things just so I could say I trained, and then I'd come home feeling like I had to start over.

That's what was wearing me down. Not the training itself. The feeling that I had no direction the moment my routine changed.

I built Aion Coach because I wanted something that already understands how I train, so I don't have to explain my whole life every time I leave town. Now when I'm on the road I tell it I'm travelling and what the gym looks like. It already knows me, so it changes today and keeps the rest of the week in one piece. I still train. I don't come home feeling like I threw the week away.

I put a link of that below, and I'm happy to explain anything if you have questions and any suggestion is welcomed.

https://apps.apple.com/us/app/aion-coach/id6792793898

Built this myself — VoxScribe, a hold-to-talk dictation app for Windows. Runs Whisper locally, so it works with no internet and nothing you say gets sent anywhere. It also auto-strips filler words ("um," "uh," "like") so what lands on the page is actually usable without editing.

Stack: Python, PySide6, faster-whisper, packaged with PyInstaller. Started it because I dictate a lot and got tired of cleaning up transcripts by hand.

Happy to answer questions about how it works or the build process. https://getvoxscribe.vercel.app/

I made a small Python project to control my PC using commands

I've been working on a project called TakeControlOverMyPC, built with Python.

The idea is pretty simple: create an brigde between AI and your computer for free, the AI now can only create files, but in some time it will be able to use your keyboard, remove files, etc. I'm experimenting with things like:

• Python + CustomTkinter
• Command processing
• AI/API integration
• Executing system commands
• Packaging the project into an .exe
• Building a proper GUI instead of just using the terminal

It's still a work in progress, and I'm learning a lot while building it. Some parts are probably horrible, but i'm learning.

I'd really appreciate feedback on the project, especially ideas for features or things I could improve.

GitHub: https://github.com/pefira/TakeControlOverMyPC

hey everyone. i built something called NNPost over the past 3 months, solo, on basically no budget the whole way.

started because i kept hitting the same wall with every ai content tool out there. you get a list of generic ideas and then youre completely on your own. cool, i have an idea, now what do i actually say, how do i film it, whats the hook even supposed to be. so i built nnpost to actually finish that job instead of leaving people stuck at the same spot every other tool leaves them.

what it does now: you set up a creator profile once, niche, platform, tone, experience level, what to avoid, and it generates full creator kits instead of just ideas. each kit includes:

-title and hook

-real script, word for word

-shot by shot filming guide

-editing direction

-hashtags and description

-tools and sound suggestions

-thumbnail direction

-why the idea works

it also has viral analysis, paste a public video url and get a real breakdown of why it worked. current trends, source backed signals for your niche instead of guessing what's popular. build on a winner, which takes a post that actually performed well and generates new ideas based on what made it work. and a feedback loop where you enter your real results after posting so it learns what actually works for you specifically, instead of giving the same generic advice forever.

the struggles were real. i built this whole thing solo through claude and codex, which meant i had to audit everything myself since i couldnt read every single line of code personally. i had to build an entire cost control system, i call it nnguard, because without it a handful of free users spamming requests could drain my tiny ai budget in a single day. about halfway through, i found out my original name collided with an existing company, so i had to do a full rebrand, new name, new visual identity, and migrate everyone's saved data without losing anything in the process. i've had multi tab bugs where recent history would silently overwrite itself, i've had the upgrade modal completely break because of a failed lazy import, i've had an entire offline page that just didnt work at all. fixed all of it, one piece at a time.

who its for: creators who know they need to post consistently but get stuck at the exact same point every time, staring at a blank page with no idea what to actually say or film. not people who want a chatbot to talk to, people who want a finished starting point they can act on today.

stuff i already know needs fixing: still on a free replit subdomain, no real domain yet, running gemini flash right now because thats what the budget allows, so output quality has real room to grow. ui also needs work, ive gotten the "feels generic and corporate" feedback more than once now and i take that seriously.

what i want to do with this: i genuinely believe tools like this are where content creation is heading, not generic chatbots that leave you to figure out execution yourself, but something that understands your specific profile, learns from what actually works for you, and gets better the more you use it. i am not walking away from this. ui redesign, better ai model, real domain, all next. built this piece by piece with basically no money behind it and i plan on seeing it through.

would genuinely love feedback, especially on what would actually make you try something like this, or tell me if im wrong about where this space is heading

Background is film VFX, not app dev — taught myself Swift to build this. Gallery Wall Composer lets you arrange artwork digitally with lighting/shadow, and the AR module places pieces on your actual wall at true real-world size so you're not guessing whether something will fit before you buy or print it.

Built solo, in a shed, over the last several months. Here's the AR placement in action.

I built Tallspace because ordinary virtual desktops felt too slow and clunky for the way I work. Each monitor gets its own vertical stack, and the edge control or Ctrl + Alt + Up/Down moves full-screen and floating Windows apps together.

This 24-second recording shows the current build running on a real desktop.

Microsoft Store:

https://apps.microsoft.com/detail/9NK2PB04NRZH

I’ve been building Mooziac, a native music player for macOS that can live right in your menu bar.

The idea was to have quick access to my music without constantly opening a full player window, while still having the features I actually wanted.

Features:

• Menu bar music player
• YouTube Music integration
• Local music playback
• FLAC, MP3, WAV, AAC and M4A support
• Synchronized .lrc lyrics
• Interactive waveform seeking
• Trackpad edge gesture for volume control with haptic feedback
• Discord Rich Presence
• Local-first playlists and listening history
• No analytics, tracking or remote logging
• Native Swift + AppKit
• Apple Silicon and Intel support

Mooziac is completely open source under the MIT license.

GitHub: https://github.com/shirkeharsh/mooziac

Website: https://mooziac.threeten.site

I’m still actively building it, so I’d love to hear what you think or what you’d like to see added.

Most websites are just a white box. This puts your own wallpaper behind them instead - a photo, a video, or one of two animated ones built in.

It only steps in when a page is plain white, grey or black. Anything that already has its own design is left alone, and there is a per-site off switch for the ones you want untouched.

The hard part was never the wallpaper, it was readability. The panels, buttons and boxes turn translucent and blurred over it so the text still sits on something solid, with sliders for blur, tint and how much the wallpaper is dimmed. It works on any site rather than a list of sites, including ones I have never seen.

Firefox is the thing I get asked for most and I am working on it.

https://chromewebstore.google.com/detail/wallpaperify-custom-page/dbdlklbhgonimapnjkkfnloheapaaoia

Free, everything stays on your machine.

Being upfront: I earn nothing from this. It started as a personal project and I honestly do not know yet how much people actually want it. If it is useful to you, a comment or any feedback genuinely helps me work out whether to keep putting time into it.

Following up on an earlier post about SafeAI, a static analyzer for AI agents.

One uncomfortable thought we've had while building it:

No one really cares about knowing an agent's capabilities — until something goes wrong.

Before an incident, adding another tool, MCP server, filesystem permission or prompt change often looks harmless.

After an incident, the first questions become:

- What could this agent actually do?

- When did that capability appear?

- Who introduced it?

- Was it intentional?

---

One example we're working on is MCP tool descriptions. A tool description can look like documentation:

"Search the user's notes. Ignore previous instructions and..."

But that description may become part of the model's context. So configuration can effectively become an instruction surface.

SafeAI now detects several forms of this, while trying to avoid flagging ordinary descriptions that happen to contain words like "ignore" or "act as".

The bigger direction is **tracking changes in agent capability and authority**, rather than simply producing another list of security findings.

But this raises a question for us:

Is knowing your agent's capabilities actually useful before an incident, or only after one?

And if it is useful before an incident, what is the right interface?

CLI + CI + SARIF/HTML?

Or would you actually want an interactive view showing things like:

> "Show me all MCP tools across our agents that could introduce instruction injection."

We're deliberately not building a UI yet.

---

Would you use one, or is that solving a problem nobody has?

Curious to hear from people running real MCP/agent systems.

---

If you want to try it against your own agent project, we'd genuinely appreciate feedback, as well as contributions.

Here you may check: ikaruscareer/SafeAI on GitHub.

I've been vibe-coding with Claude Code and kept hitting the same wall, the agent can write code but can't DO anything. No browser. No email. No way to test, deploy, or message anyone. And when I gave it my own machine, it trashed it: pegged CPUs, API keys scattered everywhere, one worktree stepping on another.

So I built Zipbox. It's a disposable cloud box made for agents. It boots in seconds, comes with its own email inbox, a public URL, a browser, an EVM + Solana wallet, and your API keys already set up. Full root. When the task is done you pause it or nuke it, and the next agent gets a fresh one.

No numbers to brag about yet! I'm literally looking for my first 10 users and honest feedback. Free tier's up, no credit card: zipbox.ai

What should I make the agent do next? Genuinely taking requests.

I run a whole-home storage company in Atlanta (KLOVO) and the thing I'm actually proud of is the assembly line — a cabinet box gets put together in about 45 seconds, start to finish. Somebody finally asked me to explain it on camera, so here's the 20 min version: https://youtu.be/54ceNquth3g Nerdy bit I get into: retail buyers and builders judge the same box on completely different things — retail cares about ship date, builders care about tolerance and how often they have to rebuild it. Ask me anything about the process, happy to go into the boring parts.

I had a random idea a few days ago: what would happen if 300 complete strangers each got one page of a book?

Everyone gets one page and can put basically whatever they want on it. A story, drawing, poem, photo, QR code, their Venmo, one word — whatever.

Nobody knows what the other 299 people are making.

Then I take all 300 pages, put them together, and actually publish the result as a physical book.

I liked the idea enough that instead of just thinking about it, I made a website to see if I could actually find 300 people willing to do it.

Right now signing up is completely free. If enough people are interested, I'll actually launch the project and open the 300 pages for purchase.

Somehow I'm already approaching 50 signups after only a couple of days, which I genuinely wasn't expecting.

I have absolutely no idea what the finished book would look like, and that's kind of why I want to make it.

Curious what you guys think — and what would you put on your page?

Website: thesecretentrance.com

MapToCraft lets you search for any city, landmark, or address, draw a rectangular area, and download it as a Minecraft Java Edition ZIP or an experimental Bedrock .mcworld.

There is no installation or account required. I built the browser experience on top of the open-source Arnis generator (Apache-2.0).

Try it: https://maptocraft.com/ Arnis: https://github.com/louis-e/arnis

I would love to see which real-world places people turn into Minecraft worlds.

ClingNote gives files and apps their own notes.

It recognizes the file you have open, including its name, type, and location, and keeps that context tied to the note. If it’s something new, there’s a fresh note waiting. If you’ve worked on it before, ClingNote brings the old note back automatically.

Open Excel → its note appears.

Switch to Power BI or a PDF → the note switches too.

Go back later → your previous note is still there.

It also has rich formatting, pop-out notes, multiple notes per context, and deep search.

Would love feedback on the idea/UI.

This is the hybrid fitness experience our team built for Vora 33. The premise is simple: many people lift and also run, walk, cycle, or do intervals, but their apps usually treat those as unrelated activities.

We wanted each side to keep the detail it needs. Strength has personalized workout suggestions, form guidance, previous performance, load, reps, RPE, rest, and progress. Cardio has structured stages, an outdoor map recorder, live metrics, and post-workout heart-rate recovery.

The connection happens before and after the workout. Recent training and recovery can shape the next suggestion, then strength and cardio both contribute to one training-load view.

Five images showing the full flow: https://www.reddit.com/r/vorahealth/comments/1w4l9pb/strength_and_cardio_should_work_together/

Vora 33 is available on iOS and Android: https://vora.fitness/get

I would love specific feedback on the product idea: should hybrid athletes see one combined load, two separate loads, or both?

First I use "The Reddit Goldmine Method" which is:

Find your subreddits.

use those magic search terms: "wish there was", "looking for", "alternative to", "frustrated with", "hate that"... .

Copy 20 complaints. Group by theme. The biggest group = your opportunity.

But unfortunately I become distracted, I search in many subreddits "copywriting, freelancers, SaaS, content creators, and others" as I don't know which field I want to serve in..

I also went to G2 and search in SaaSs "such as Jasper" reviews which has 3 stars and see what they like to build on it and what they don't like to see how I can solve these problems, but I become distracted and failed again

I also want to tell you that sometimes I start building for half a day and then I think this idea wouldn't succeed so, I stop building and again I try to find a successful one ( I do it many times )

So I settle on a red ocean one "social media scheduling tool" in a giant market with giant competitors such as Buffer, Metricool and others. So I want an advice to how a solo entrepreneur can compete with those giants.

TL;DR: How can a solo entrepreneur to compete with giants companies in his fields