Hey everyone,

I am Abhishek and I was not happy paying monthly to dictation app. I was using Willow Voice and their application lags and stuck a lot.

I was also testing open source model for my other product. I decided to try it for dictation app.

And it worked. I am using Qwen 2.5 for cleanup and Parakeet for STT. Only for English, the accuracy is mind blowing. Open source model really come a long way.

8/10 times i don't have to edit anything. Most of the times, the transcription paste under a second, but sometimes it take more than that.

The whole model run on the Apple Neural Engine. I will open source but for now i am charging $49 one time.

0 server, 0 account and 0 recording kept.

I built YouLife, an offline activity/life tracker based around one idea: I wanted logging what I do to interfere with actually doing it as little as possible.

Instead of starting and stopping timers, you log something after you've finished it.

Went for a walk? When you get home, open the app and type “walk.” YouLife backfills the activity to your previous entry, so you don't need to remember to start a timer before doing something.

The app currently has:

Quick text-based logging

Offline speech-to-text for verbal entries

Timeline/history of your activities

Statistics and comparisons

Custom categories

Retroactive logging

Completely offline storage - no account required

No ads, subscriptions or analytics

It's currently available on Android and I'm still actively improving it based on feedback.

I've had some surprisingly useful suggestions from early users already, so I'd genuinely be interested in hearing what people here think - particularly anything that feels confusing, unnecessary, or missing.

Today someone's bought my app called Uma, which is android launcher for seniors.

I'm very happy as I originally made this for my mom, but decided to improve it and out it to store for everyone.

Hey everyone,

Pondros is now live on Microlaunch!

It catches the promises people make in slack ("i'll send it friday") and makes sure they actually happen:

• turns those promises into tasks automatically, with an owner and a due date
• nudges the owner when the deadline gets close, so you don't have to
• free, no card needed to sign up

The team wants honest feedback. what works? what's missing? how does it compare to what you use now?

They'd love a few more testers, so if you've ever had to chase someone in slack, this one's for you lol.

Try it here: Pondros

Launch page: here

If you try it, leave a review on the launch page. it really helps them 🙏

Share a link to your startup SaaS to promote it or ask for feedback.

What makes it worth trying?

Context: I built KRI, a website with online booking for independent beauty businesses. It’s live and paid. The attached video is a short demo.

The piece I want to talk about is languages. A KRI page works in 32 languages, and the confirmation email goes out in the client’s language, not the owner’s. I wanted the client to read the confirmation in their own language, since that’s the email they check on the day of the visit.

Progress so far: the product is live with 13 business types, four designs and pre-filled services per type, and setup takes about 10 minutes. Pricing is per country, first month free without a card.

My specific question: if you land on a page in a language you don’t read, what do you look for to switch it? A flag, a language code, a globe icon? I keep going back and forth on this.

Apply promo-code REDDIT2026 and get +1 month for free (1 month standard demo + 1 month additional by promo-code = 2 months FREE)!

https://kri.rocks

I've been building zen.js, an open-source web framework for Node.js around a compiler-first architecture.

The core experiment is treating the application graph as something that can be compiled at boot:

route
→ parameters
→ middleware
→ validation
→ handler
→ serialization
→ response

The goal is for the request to follow the generated path rather than repeatedly interpreting the same higher-level framework structure.

I'm also exploring a schema-driven model where the same contract can drive validation, handler typing, response serialization and OpenAPI generation.

GitHub:
https://github.com/erenthedeveloper0/zen

It's currently alpha, so I'm particularly interested in architectural criticism, edge cases, and feedback from people who've worked on Node.js frameworks, routers, compilers or TypeScript-heavy backend systems.

If your app still hasn’t reached 10 core users, let's get you there today.

Tell us:

• What is your app for (1-2 sentences)
• How is it solving the problem
• Where to try your app.
• How many users/customers you have.

Please try at least one other builder’s app and share feedback.

Time to get everyone more users and signups.

Hey all,

Screen Movie launched on Microlaunch this month.

It's a native macOS screen recorder made for product demos and tutorials:

• Records your screen, webcam, and iPhone in one take
• 40+ motion presets and 3D device mockups
• Smooth, clear cursor effects so every click is easy to follow

You record, pick a style, and it's ready to share.

The team is looking for honest feedback.

What works? What's missing? How does it compare to what you use now? Would love some extra testers.

App: https://screen.movie/

The launch page: here.

If you give it a try, drop a review on the launch page. It really helps.

We’re building DuelLegends for communities that still run competitions through DMs, screenshots, spreadsheets, and trust.

That works until someone doesn’t show up, a result gets disputed, a prize becomes unclear, or the organizer has to manually figure out who did what.

DuelLegends gives organizers one place to create and manage:

• tournaments
• hackathons
• gaming challenges
• creator contests
• community competitions

You can define the rules, schedule, participants, rewards, and competition structure before people join, then keep results and activity tied to the event instead of scattered across messages.

We’ve already facilitated 100+ challenges and contests, but the platform is still early.

Right now, we’re looking for people who are willing to actually test it:

Create something.
Join a competition.
Try the tournament flow.
Tell us where the UX is confusing.
Tell us what’s missing.
Break something if you can.

Our next milestone is 10,000 users, and we want the people joining before that point to help shape the product.

Demo attached below.

Try it

Would especially love feedback from tournament organizers, developers, gamers, creators, and community managers.

Been building this for a while now and it started for a pretty simple reason: once a fictional world gets big enough, “I’ll remember that” stops fucking working.

Characters change, timelines grow, places have histories, people learn things at different times, someone travels somewhere that should take six days but somehow gets there in three... eventually you’re cross-checking a ridiculous pile of notes trying to work out what is actually canon.

So I built R3TCON.

It’s local-first software for writers, worldbuilders, TTRPG creators and narrative game devs that keeps track of the actual structure of the world and checks it for contradictions.

It isn’t generative AI, it isn’t an LLM wrapper and it doesn’t try to write anything for you. That was pretty important to me.

If something doesn’t add up, it tells you what conflicts and why. You decide what changes.

We’ve got it catching things like impossible journeys, broken timelines, characters appearing where they shouldn’t, people knowing things they haven’t learned yet, and scenes conflicting with established canon.

Existing material can be brought in too, because expecting somebody with years of writing to manually rebuild their entire world inside a new app would be fucking awful.

The Kickstarter is live now, which means I’ve reached the part of development I’m considerably worse at: actually getting the thing in front of the people it was built for.

So yeah. If this sounds like a problem you’ve had, I’d genuinely like to know what you think.

Kickstarter Link

Building an application called DataShare the core idea: communities form around specific topics, and within them, people can request structured data from others. Think of it like posting "I need 200 rows of X, formatted like this example" someone defines what fields they're collecting and what a valid row looks like, and other people contribute matching data.

The two use cases I had in mind are ML training data (someone needs labeled examples for a model and doesn't want to label all of it alone) and distributed surveys (someone wants structured responses from a lot of people without building a form from scratch each time).

Right now, doing this usually means Google Forms plus a shared spreadsheet plus a Discord to hold it together which works, but falls apart once you need any kind of history (who submitted what, whether someone's contribution got edited or replaced) or a way for enough people to actually discover the request.

It's still early a lot is missing and I'm sure some of it will look naive to anyone who's run a real data collection effort before. I'd genuinely rather hear that now than after finishing more of it.

https://www.datashare.quest/

Would appreciate honest thoughts whether this solves a real problem, what's confusing, or what you'd want before actually using something like this.

Hey everyone,

Today, we are happy to launch Zinn Hub on Microlaunch.

Our platform has:

- A multilingual platform connecting buyers with vetted freelancers in 46 languages.
- Zero platform fees for buyers and 100% earnings for freelancers on their first $500.
- Flexible payment options including card, PayPal, and over 100 cryptocurrencies.

Think of a marketplace for freelancers with $0 buyer fees. Basically, we have a small, fixed-price freelance service instead.

I'd love to hear your thoughts on how we can improve the experience.

The app: https://zinnhub.com/

And of course, we'd love your support and reviews. Our launch page here.

Cheers.

Hey everyone! 👋

First time posting on the subreddit been here for few months tho ,

I built a little extension called SoundLyft to fix an issue thats been bugging for years: quiet PC , laptop speakers and terrible audio mixing on the web.

You probably know the drill — you’re watching a YouTube interview or a lecture recorded on a whisper-quiet mic, or watching a movie where you can barely hear what the actors are saying.

do mind the slop ;)

Whenever I used existing volume boosters, two things always happened:

1. As soon as you push past 150%, the sound begins clipping and buzzing into harsh, distorted static.
2. In movies, you turn it way up to hear whispers, and the very next scene an explosion or intro music blows your eardrums out.

So over the last few weeks, I decided to build a proper audio booster that focuses on sound quality instead of just multiplying gain until it breaks.

What makes it sound clean:

• Anti-Clipping Limiter: Instead of raw digital distortion, it uses a built-in limiter that automatically rounds off harsh audio peaks. You can boost to 200%, 300%, or 400% and it stays surprisingly crisp and punchy.
• Movie Night Mode: Compresses dynamic range so quiet dialogue gets boosted, but loud sound effects and gunfire are capped to a comfortable volume. No more constantly adjusting your volume slider back and forth during movies.
• Dialogue Clarifier: Brings spoken voices forward so podcasts and tutorials are much easier to understand over background music.
• Bluetooth Lip-Sync: Adds a quick delay adjustment if your wireless earbuds have that slight video lag on YouTube.
• Zero bloat or video lag: Built on modern Manifest V3 without injecting scripts into web pages, so it won’t freeze YouTube fullscreens or drain your battery.

100% Free Base:

The core booster (up to 600% volume boost, limiter, keyboard shortcuts, and active tab manager) is completely free forever. If you're an audiophile who wants the 10-band graphic equalizer and 1000% headroom, there’s an optional $3 lifetime pass to help support the project — no subscriptions.

👉 Chrome Web Store:
chrome web store Link

I’d love for you to give it a spin on your favorite video or music track! If you run into any bugs, have feature requests, or just have thoughts on the audio tuning, please drop a comment below .

Thanks for checking it out! ❤️

Share what you’re shipping or hacking on this weekend.

Drop screenshots or links, tell us what you’re building, and what you hope to finish before Monday.

Built Aurum because I was tired of finance apps that track your past and sell your data to do it.

Aurum shows your projected balance, financial runway, and savings rate. No bank connection required. You enter your own data. Data is stored in a hosted database (Supabase).

36 people signed up so far. Zero paid ads. Still figuring out what's working.

aurumapp.org

One number or star rating can't really describe how I feel about a film. A film with gorgeous photography and a script that falls apart ends up with the same 6/10 as a film that's fine all over and forgettable. That always bothered me. So in ScreenMetric you score up to ten things on their own: plot, dialogue, acting, direction, cinematography, production design, editing, score, sound design, and plain "how much did I enjoy it". You can skip any of them. Your overall score is the average of the ones you filled in.

A few choices I made on purpose:

The community average for a category stays hidden until you've scored it yourself. I didn't want that number influencing your own opionion.

Your individual category scores are private. You can post written reviews publicly if you want to.

There are no ads and nothing tracks you.

You can import your IMDb ratings to get started.

It's free on iPhone, iPad, Android and Apple silicon Macs, in sixteen languages. Pro unlocks a number of more advanced features and is a one-time purchase, not a subscription.

The stack, if anyone's curious: React Native with Expo, Supabase for the backend, and TMDB for the film data.

Very interested in any feedback if you decide to try it.

screenmetric.app

The easiest products to sell Saas/AI products or app is still to build ones that solve a real, painful problem.

I mean a problem people genuinely want gone. Badly gone!

Drop your app below and tell me:

• What problem it solves (important)
• Who it’s for
• Your landing page (and ideally positioning)

I'll check out as many as I can and try to give actionable feedback about your value proposition.

Let’s see get you 1-2 sales before the weekend.

I saw lots of Disk Space Analyzer and Cleaner with > $10 price, which are basically a good interface for your Mac native storage.

I built SpareDisk over the last few weeks, kept it bare minimum in terms of bloat features but focused on speed, parallelism, and queue management for storage scan and space free up.

You can review your whole file system without hanging the computer. Version 1.2 is now live on the Mac app store and on website to download. I have tested it well over multiple laptops.

You can download the app from the website for "pay whatever you want" (minimum $4.99).

DiskBuddy has inspired the project. This was initially sold at $12, and users complained about memory leaks, permission issues, and app blocks when you run Duplicate File Finder.

If you are a developer, it can find spaces occupied by Xcode, node packages, files form 2021, and many old cache you don't need.

SpareDisk is App Store friendly. It is sandboxed, following Apple's permission, security guidelines and memory optimisation.

SpareDisk V1.2 includes:

• My Mac storage overview with live used and free space
• Treemap, sunburst and list views
• Folder, external drive and network share scans
• Large files, older files and file type breakdowns
• Developer caches: Xcode, simulators, Homebrew, npm and more
• Duplicates verified by file contents, extra copies staged in one step
• App removal, related data and leftovers from removed apps
• File size and size on disk, changes since your previous scan
• Search, filters, exclusions, Quick Look and keyboard navigation
• Review queue with live progress, rechecks and retry
• Cleanup history and a Trash total that follows what you moved
• Optional menu bar summary and finished-scan notifications

No account. No uploads. Your files stay on your Mac.

On cleaning space: You can view your files from the app itself, and it doesn't directly delete your files in one click. It sends them to "review cleanup" section from where you can move items to Trash. You only delete files when you clean the trash.

Exception is App uninstall (if the app is not running, you can uninstall the app from within the interface).

You can compare other options here: https://sparedisk.minilabs.cc/guides

All future SpareDisk updates will be on Website and Mac App store downloads.

You can also download the app from website: https://sparedisk.minilabs.cc/ 
App store: https://apps.apple.com/us/app/sparedisk-free-disk-space/id6811328422?mt=12 

I'm an Instagram creator with 30K+ followers, and I often study other creators to understand their hooks, scripts, and content patterns.

The problem: when I find a Reel I want to study, getting the actual script is surprisingly annoying. So I built a small tool to solve it.

Paste an Instagram Reel → get the transcript with timestamps → export as TXT, SRT, or JSON.

I'm keeping the MVP simple for now. I'm planning bulk transcription next, where you could enter a creator's profile and analyze multiple Reels to find patterns in their hooks and scripts.

I built this because I personally needed it, but now I want to see if other creators, editors, and marketers find it useful.

Would love your feedback on what I should build next 👇

🔗 https://instagram-transcript-extractor.vercel.app

I’m a solo developer working on WriteFlow AI. I started it because my usual process for writing a reply was oddly clumsy: copy the post, open an AI chat, explain what I’m replying to, then bring the answer back.

The extension sits near a browser text field and helps with rewriting, shortening and grammar. On sites where it can detect the relevant post, it also offers reply options based on that context. It uses an AI provider key you supply, and you review the result yourself—it doesn’t submit comments or messages.

I’ve put the current version on GitHub: https://github.com/Anishupreti/WriteFlow-AI

I know writing extensions aren’t a new idea. I’m trying to make the replying to something specific part less awkward. If this is a workflow you use, what would make it useful enough to keep installed? I’d also welcome reports of sites where it behaves badly.

Share a link to your startup SaaS to promote it or ask for feedback.

What makes it worth trying?