PH热榜 | 2026-03-21

Hi Product Hunt,

I’m Mu, founder of Lokuma.

Before this, I built Readdy and Creatie — tools used by over 500,000 designers and creators. Most of my work has been around design systems and how products actually feel, not just how they function.

This time, I’m working with a small indie team — a mix of designers, AI researchers, and people from growth and marketing.

Recently, something started to shift.

Tools like Claude Code, Cursor, Codex, and even things like Google Stitch, are changing how software gets built.

We’re no longer just using tools ourselves.

We’re starting to work with AI, and AI is starting to use tools on our behalf.

That changes the interface.

The new “user” of many tools is no longer a human - it’s an agent.

And most tools today aren’t built for that.

AI can generate almost anything.
But generation isn’t design.

What we see today, including a lot of outputs from varied AI agents, is that AI can produce UI, code, layouts very quickly.
But the results often lack structure, hierarchy, and visual coherence.

They work.
But they don’t feel designed.

That gap becomes much more obvious when AI is the one building.

So we started thinking:

If agents are going to build products,
they need something closer to a designer, instead of just another generator.

That’s why we built Lokuma Design Agent.

It’s a design intelligence layer for AI - something your agents can call to reason about layout, typography, hierarchy, and visual balance.

Your AI handles logic and generation.
Lokuma handles how it actually comes together.

Why now?

Because we’re entering an agent-first stack.
Coding has agents. Research has agents. Execution has agents.

Design is still missing.

Why us?

Because our team brings together experience across design tools, real product building, AI systems, and growth.

We’ve spent years understanding what makes interfaces actually feel right, and now we’re translating that into something AI can use.

We’re still early, just a small indie team moving fast. But this feels like a new category:
tools built not for humans, but for AI.

Curious how others here see it:

If your AI is already writing code,
what’s still missing for it to ship something that feels complete?

— Mu

I've been part of the marketing team on two AI website builder tools. We watched the category explode and then hit a ceiling. The category grew fast, but the outputs started converging. Generic layouts, same patterns. Hard to keep up with what actually feels current. AI got really good at generating. It never got a good sense for design.

So when Lokuma decided not to build "yet another AI website builder" and instead focus on being the design layer for AI agents, I thought that was the right call. Crowded category and it really was the yesterday's problem.

It works as a skill that your AI agent can directly call. You just tell Claude Code or Codex to " install and design with Lokuma," and it actually reasons through the layout before generate anything.

It's still early, and to be honestly there's a lot Lokuma can't do yet. But the overall direction feels right to our team. Your agent handles the logic, Lokuma handles how it actually comes together. Hierarchy, balance, what guides the eye.

Hello everyone, I'm Joy, co-founder of Lokuma.

Before starting Lokuma, I was running a branding and marketing agency focused on helping SMBs build their online presence covering social media, branding, and website development. But over time, I noticed a recurring problem: many SMBs were stuck with either an outdated website or no website at all, simply due to limited budgets and resources.

But things are shifting. I saw many entrepreneurs now having AI agents such as OpenClaw, Claude Code to help them. And this is why we're building Lokuma Agent. Simply install our agent with any of yours be it Claude Code, Cursor, Codex and watch your own agent design website in seconds using Lokuma's intelligence.

This will really help individuals to offload heavy lifting tasks to our agents and equip your personal AI agent with website development intelligence to build websites that TRULY knows design & conversion logic!

This is a groundbreaking function in the world of AI agents and we have received many positive feedback.

We'd love to have you to try it and offer us any feedback!

been running openclaw for a week now and the one thing missing was decent UI for the agents' output. this fills that gap nicely. does it support custom themes per agent?

Been building stuff with Claude Code lately and the design gap is real. Everything comes out looking like a bootstrap template from 2018 lol. The idea of plugging in a design layer instead of manually fixing spacing and typography every single time is exactly what I need. Gonna test this out today.

This is a really interesting direction.

AI can generate fast, but “good design” is still the missing layer — structure, spacing, hierarchy, all the things that make something actually usable.

I like how Lokuma positions itself as a design intelligence layer for agents instead of just another generator. If this works well in practice, it could become a core piece of the agent stack.

Curious to see more before/after examples — that would make the value even clearer.

it’s time to get SMB really powerful design agent! Congrats Mu and Joy launching this new product!

Hey everyone! Tech Lead at Lokuma here. 🛠️

We built Lokuma because we were tired of AI-generated websites that looked like templates from 2010. Design Agent provides a sophisticated design layer that any agent can call via a simple API.

Works with: OpenClaw, Claude Code, Cursor, and more.
Does: Aesthetic reasoning, typography, and visual balance.

We’re excited to see how you integrate this into your agentic workflows. I’m here for any questions on the tech stack or our future roadmap. Let’s make the AI-generated web beautiful! ✨

as someone who don’t know design, his product helped me a lot

Really strong positioning.

AI can generate fast, but making outputs feel clear, usable, and visually refined is still hard.

Love that you’re focusing on the design intelligence layer instead of just more generation.

Congrats on the launch and upvoted.

Would’ve loved to have this a few projects ago!

I’m not surprised this came from Mu’s team. They’ve always been sharp on product feel, not just functionality.

Huge congrats to Mu Li and the team! This is very impressive work, and I can't wait to try it soon. Well done!

Well done @mu_li !! It moves fast without sacrificing craft.

Most AI outputs feel like rough drafts. Lokuma feels closer to something you’d actually ship. The structure, spacing, and visual decisions make a real difference.

Feels like giving AI a sense of taste.

Congrats on launch. Just a friendly suggestion, but it triggered my OCD hahaa, please center "Random Typography" vertically on: https://agent.lokuma.ai/group/group-2-before.png

Very Very cool! let me try it today! I really need a designer to uplift my sites right now :P

Congratulations on the launch! I truly believe that design is a weak spot of today’s agents — happy to try Lokuma in action.

Typography is surprisingly good here.

This is very good product, I'm eagerly waiting to try

On Design Agent by Lokuma, the generation-vs-design framing is the part that lands. Agent outputs that technically work but don't hold together visually is exactly where trust breaks down with AI-built products. If Lokuma can enforce design system consistency across outputs from different agents - Claude Code, Codex, Cursor - that's the hard problem worth solving.

the rare combination of taste and tech. well done lokuma team!

Super interesting! I’m going to check this out!

Congrats on the launch!

Finally, a designer for my agents! I use OpenClaw and Codex daily, and having a way for them to actually reason about hierarchy and balance instead of just spitting out generic components is a game changer. Huge congrats to the team for identifying this missing piece!

This fits the moment.

Really like this.

It feels pretty clear at this point that agents are becoming a real product layer, not just a feature. We’ve been thinking about the same shift at Lokuma from the design side — if agents are going to change how products get built, then design can’t stay as a separate, manual step in the old workflow.

That’s a big part of how we think about Lokuma as a design agent — a system built to turning intent into actual creation in a way that feels native to this new era.

Great launch.

Claude Cowork Projects turns Claude Desktop into a structured workspace for ongoing AI tasks.

Instead of one-off chats, you can organize work into project-based environments with local folders, instructions, memory, and scheduled tasks solving the chaos of fragmented workflows.

What makes it stand out is its local-first, project-scoped approach with persistent context and memory, plus recurring task automation inside each project.

Key features include project-specific instructions, memory, scheduled tasks, and folder-based context.

Perfect for researchers, operators, and teams running repeatable workflows like reporting, analysis, or knowledge work.

P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

Yes, finally there is my dream come true. Because this was for meh the mail issue.

The issue I have with Cowork and also with local projects is, that my local file system is a mess. I have several versions of the same file and it's hard for anyone who is not me, to figure out which is the most current version. While in the usual online projects, I work clean and always delete and replace the most current one. It's not just file versions, but also the overall focus that's different. I still try to find a way to fit Cowork into my workflow and figuring out what the advantage is compared to online projects.

the context management is the part that matters most. i run a full business off claude code and the biggest friction is keeping project context across sessions. how does this handle large codebases?

Having consistent context across tasks is the part that really resonates. Right now I juggle between so many tools and lose context constantly. The local-first approach is a nice touch too.

Anyone noticed the video recording doesn't actually work. It shows an error - "Due to capacity constraints..."

The local-first approach with persistent context is the right call — the biggest pain with AI workflows isn't capability, it's continuity. Every session restart wipes your mental model. I've been building in this space and context bleed between tasks is the #1 productivity killer. What I'd love to see: explicit context scoping per task (not just per project), so you can run parallel subtasks without them leaking into each other. Also curious how you handle scheduled task failures — does it surface errors with enough context to recover without starting over?

Really interesting concept.

I noticed some UX friction around project structure and how users understand the difference from standard Claude workflows.

Happy to share a quick teardown with suggestions if useful.

Having Vite, Vitest, Oxlint, and Oxfmt all unified under one CLI is huge. Right now my projects have separate configs for ESLint, Prettier, Vitest, and the build tool – and keeping them in sync is easily 10% of my setup time for any new project.

One thing I'm curious about: how does `vp check` handle framework-specific linting rules? For example, React has its own set of rules (hooks rules, JSX accessibility), and Vue has theirs. Does Oxlint cover those out of the box, or would you still need to layer on additional plugins for framework-specific checks?

Runtime + package manager + frontend stack in one entry point is a bold scope. Where does Vite+ draw the line - is the goal to replace something like Volta or fnm for runtime management, or is it more of an opinionated layer on top of them?

Oh this is nice. Setting up a new project always feels like I'm spending the first hour just picking and configuring tools instead of actually building. If this actually unifies all that into one step I'm sold. Starred the repo.

switched to vite last year and never looked back. curious what the + adds — is it mainly the unified config or are there perf wins too?

:) Been a Vite user for a long time. Hope this one delivers too

Really interesting direction.

I noticed some potential friction around how new users understand the scope and differences vs existing tools (like Vite setup, linting, etc).

Happy to share a quick teardown with suggestions if helpful.

Really like seeing Evan You go all-in on the unified toolchain approach. As someone who's wasted too many hours wrangling ESLint + Prettier + Vitest configs on every new project, having one CLI that handles all of it is exactly what the ecosystem needs. The Oxlint/Oxfmt integration is the cherry on top — Rust-based tooling speed with zero config overhead.

Hey Product Hunt 👋 

I'm Hanh, co-founder of Fractal.

We were early participants in the ChatGPT Apps working group when we noticed a shift: people don’t just ask AI questions, personal AI agent (like ChatGPT, Claude, etc) are becoming the main interface for getting work done, and distribution is moving from the web to the AI agent people use every day. That is why ChatGPT apps matter. If your product is not something the agent can discover and use, you are missing out on a distribution opportunity.

But ChatGPT apps aren't just web apps inside an iframe. Great apps are:

• Context-aware: take into account conversation history and user memory

• Action-taking: translate intent into actions

• UI-augmented: easy for both human & AI to interact with

So we built Fractal, a platform specifically for building outstanding ChatGPT apps:

• Plan: a smart architecture planner that recommends best practices, including what belongs in the app vs what to delegate to the model

• Build: a specialized AI coding agent that understands context, actions, and UI constraints

• Iterate: a native chat emulator so you can test without reconnecting to ChatGPT every time

• Ship: one-click deploy

We are the Lovable of ChatGPT apps: in the AI era, it's not enough to be lovable, you need to be frac-able 😉

We’d love your honest feedback, and we’ll be around all day answering questions.

🎁 Product Hunt launch offer: 20% off Pro with code PH20

👉 Join our Discord: https://discord.gg/PyKaykcH

This is really interesting, how customizable are the apps we can make with Fractal? Can we implement custom code and more heavy duty applications as well?

Interesting idea. How do you balance speed with reliability when Fractal is handling everything from planning to deployment? For example, how much control do developers have over the architecture and code quality, and how easy is it to step in, modify, or debug parts of the system without fighting the abstraction?

P. S. Good luck!

I've been building a couple of chatgpt apps too - excited to give this a try!

Love that video in the style of Wallace & Gromit! :D

I have tried Fractal for a few days and have had a nice experience so far! It is very easy to navigate and the vibecoding is pretty robust. Hopefully there will be more options for customization to work with the vibecode.

There can still be some risks relying on the ChatGPT platform alone, but I think there is great potential in this space to grow.

Anyways, nice product!

Love the focus on speed + iteration, feels like that’s the real edge right now.

Curious how you think about platform risk though, if everything lives inside ChatGPT, how do builders hedge against that long-term?

Congrats @hanh_nguyen !! Fingers crossed 🤞😼

So lucky to get to try this early on. Loving it so far!

@hanh_nguyen how is it different from codex?

@hanh_nguyen how is it different from codex?

@hanh_nguyen how is it different from codex?

speed to ship is everything right now. built a saas mvp in 4 days with claude code — tools like this that cut the boilerplate are exactly what solo builders need.

Very cool congrats @hanh_nguyen

the best no code chatgpt app builder out there - we've been using fractal since the early days and its been amazing for us to both quickly ship apps and also prototype and test new features with our enterprise clients. congrats on the launch team!

Cursor Glass is a new interface for agentic development that solves a real pain: managing multiple agents across local and cloud environments without losing visibility.

With a unified workspace for agents, repos, and tasks plus Cloud Handoff to switch between local and cloud mid-task, it reduces context-switching and keeps workflows seamless. Composer 2 adds strong performance at lower cost.

Built for engineers running parallel agents at scale especially useful for multi-agent workflows and cloud orchestration.

P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

This is a great release by Cursor Team. So, this is actually filling the gap where all the other models were doing really well, and I tried the Composer 2. It's fantastic. The speed is amazingly fast.Kudos to the team and well done.

What's the approach you're taking to ensure consistent performance and seamless handoff between local and cloud environments, especially considering potential latency or network issues?

Hey Product Hunt 👋

I'm Devansh, 18 year old engineering student from India. Me and my co-founder Samagra built Contral over the last 6 months with zero funding.

The problem we kept hitting: we were vibecoding entire projects and couldn't explain any of it in interviews. The AI made every decision and never explained a single one.

So we built the fix directly into the IDE.

Contral teaches you every line while AI codes your project. Not after. Not in docs. Right as it happens.

Key features:

→ AI writes the code. You learn while it does.

→ Teaches every function and architectural decision in real time

→ Codebase Analyzer: tested on a 10M+ line repo, mapped full architecture and generated quizzes from actual production code
→ Learn Mode: structured curriculum starting with Java

Download today at contral.ai/download

Would love your honest feedback, brutal is fine, we can handle it. Ask us anything 🙏

I'm so happy I found your product! I'm a solo founder and built my product entirely with the help of AI. At first, I watched a ton of tutorials on YouTube and even took a paid course, but I didn't really get it because you have to try everythin out in practice right away. Eventually, I managed to figure out GSD, and it built the app for me. But I’d be lying if I said it went even halfway smoothly. No, it constantly forgot what we were doing, mixed up tasks, and just didn’t understand me. And I didn’t understand it. I still don’t. I’d love to test your product.

@samagra_gune Really interesting direction, especially the “learn while building” inside the IDE itself.

One thing I noticed from the discussion here is that people are slightly confused about how Contral is fundamentally different from existing tools like Cursor or other AI coding assistants.

Have you thought about making that difference more obvious visually or through a simple before/after use case?

Feels like once that clarity clicks, the value would be much easier to grasp instantly.

How does the Learn Mode adapt when you're already familiar with a language but just new to a specific framework? Congrats on the launch!

Congrats on the launch Samagra! quick questsion, how does Build Mode decide when to step in with assistance versus letting you work through something on your own?

this is such a smart approach — most AI coding tools just do the work for you, but Contral actually helps you understand what's happening while you build. the Learn Mode concept is brilliant, especially for developers who want to level up without sitting through tutorials. the fact that 200+ devs are already on the waitlist before launch says a lot. excited to see where this goes, congrats on the launch!

But if we create a rules file for the ide we use ( ex : Cursor ) , it remembers our choice of getting the code explained to us , so after every code / file it make changes , the AI itself would write an explanatory blog or technical muse for our understanding.
just for serving a purpose of learning while vibe coding ~ this can be get done by existing IDEs & AI plugins... i am still not convinced abt the product fit for the problem you are solving, cuz its a solved problem.

Writing this cuz you asked for a genuine review , besides that - forking a open source IDE and adding AI integration is a skillful task tbh , good job !

Hey Product Hunt!

I grew up texting. My generation booked things online, ordered through apps, and found every possible way around talking to strangers on the phone. So most of us never really learned how to be comfortable making calls.

Not the "I hate small talk" kind of uncomfortable - more the kind where you rehearse what to say to a doctor's receptionist for twenty minutes, then hope they don't pick up. Job hunting made it worse: phone screens, salary conversations, follow-up calls - each one felt like a test I hadn't studied for.

So I built murmur: an iOS app that lets you practise stressful calls with AI before making them for real. The AI plays the other person with a real voice, and gives you feedback after. Doctor's appointments, landlord negotiations, awkward returns, mock interviews, salary talks - and custom scenarios if you have a specific call you're dreading. The AI shows emotions, hesitates, pushes back, and tries to behave like an actual person on the other end.

Murmur got featured in the App Store's "Best New Apps and Updates" a few days ago, and it's for anyone who has ever made a phone call in their head before making it for real. Would love to hear what call you'd practise first.

@valzevul This is super relatable — the “hoping they don’t pick up” part hit hard 😄

Really like the idea of practicing real conversations with an AI that actually pushes back. One thing I was wondering — have you explored showing a more “live” feel of the experience on the landing?

Like a quick simulated call flow or visual example of how a conversation evolves could make the value click instantly for first-time users.

Feels like once people *see* the experience, they’d be much more likely to try it.

Congrats on the launch! The "rehearse before you call" concept is something a lot of people need, but I believe nobody has built well until now? At least, for me, it's the first time seeing it like this - great work!

Curious whether you've considered business scenarios like investor calls or client negotiations. At CoreSight, we work with founders who'd probably use this before a fundraise.

Hello!
I tried murmur and I have some notes,
1.It was very hard to find the app on the appstore, I had to search for murmur ai because lots of apps had the same name, maybe you wanna change it to murmur practice or something that helps users get there faster
2.In my case i was not able to test it properly as it directed me to the demo and later on it asked me to make a payment, I was not able to test it in any way, i believe is hard to get users to pay if you dont give them even a small sneak peak.

3.UX: I feel its very closed, I cant access anything , cannot setup my language preferences, speed of speech anything (maybe this changes with the premium features but im not sure)

would love to test it out though!

I'd actually use this for practicing calls in a second language. When I was in the US, making phone calls in English for things like appointments or reservations was always way more stressful than it needed to be. Being able to practice the exact conversation beforehand would have helped a lot. Does it support other languages like Japanese, or is it English only for now?

This is such a clever use case for voice AI. The "rehearse in your head before dialing" thing is so universal — turning that into an actual practice session with feedback is brilliant. Love that the AI pushes back and shows emotions instead of just being agreeable. That's what actually builds confidence for the real call.

The "hopes they don't pick up" thing is so real. What makes this interesting to me is whether the AI actually pushes back - most roleplay tools are too agreeable and you walk away feeling more confident than you should. Does murmur simulate the person being annoyed, distracted, or just saying no outright? That is the scenario that actually trips people up.

This hit close to home, Vadim 😄 I still rehearse "hi, is this a good time?" before calling my own bank. The part about hoping they don't pick up — way too real.

The idea of AI that actually pushes back and hesitates is what makes this special. Most practice tools feel like talking to a wall. This sounds like it actually builds real confidence.

Congrats on the App Store feature! Trying this before my next client negotiation call for sure 🚀

Replit Agent 4 is an AI-powered builder that lets you design, build, and ship apps, websites, and more... all in one place.

Building software today is fragmented, slow, and full of coordination overhead. Agent 4 solves this by handling execution in the background so you can stay in creative flow.

What makes it different is the unified environment + parallel agents. You can design on an infinite canvas while the agent builds, run multiple tasks at once, and ship everything from apps to decks, inside a single project.

Key features:

• Infinite design canvas with UI variant generation 🎨

• Parallel task execution with automatic conflict resolution ⚡

• Smart task sequencing submit requests in any order 📋

• Web apps, mobile apps, decks, data apps from one project 📱

• Integrations with Linear, Notion, Databricks, and more 🔗

The benefit: faster iteration, less context switching, and way more progress per minute.

Perfect for builders, PMs, teams, and anyone who wants to go from idea to production faster. If you’re building anything, this is worth trying.

P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

The "one flow" pitch makes sense for greenfield builds but I am curious how it handles mid-build pivots - when you are 60% done and realize the data model needs to change. That is where most AI builders I have tried fall apart, they are great at starting but bad at the messy middle. Does Agent 4 have a way to reason about what exists before making changes, or does it mostly rebuild?

I really liked the idea, it's a truly successful initiative, congratulations!

Curious what happens with "okay" over time - does it tend to drift toward regret or worth it as people use the app longer? That middle category feels like where the most interesting behavioral data would live.

Congrats on the launch! The "how did that feel?" approach is way smarter than categories - it's honest.

Does it show you patterns over time? Like "you regret coffee shop purchases but not restaurant dinners" kind of thing?

Hello Alex!
I hate the feeling "where did my money go?" and I have noticed when I pay cash/transfer, I am more responsible, while when im paying with credit card im extremely irresponsible. So I was thinking it would be great to be able to track the method of payment, so if i really see a trend with my CC I would start leaving it at home!.
Besides that I would like to add another choice on the feelings like "Needed it" there are some things I hate to pay but its necessary, not like a subcategory but an expense that is a non negotiable in my life and cant be decreased (or changed), for example my student loan monthly fee, rent, insurance , etc. I honestly dont feel at my best paying it, but its a non negotiable.

Sounds like a great idea to solve overspending, and it would be nice for it to sync with budget trackers people already may use (e.g. Notion) for bigger picture budgeting / factoring in necessary spending such as rent and utility bills. How does Mindspend know your transactions, do you have to manually input them?

This is a really interesting angle on budgeting. I'd love to see this go further. Imagine if it could integrate with e-commerce platforms and show how other buyers felt about the same product over time. Like an emotional review layer on top of your purchase history. With sites like Amazon flooded with fake reviews, something like this could become a genuinely trustworthy way to judge whether a purchase is worth it.

Really interesting angle — most budgeting apps make you feel guilty about spending. Focusing on how it felt instead of just the category is a much more honest way to understand your habits. Congrats on the launch 👍

Hey everyone! 👋 I'm Hojong, the solo developer behind Novi Note.

Why another note app?

I'm a backend developer working across TypeScript, Kotlin, and Swift every day — jumping between different IDEs and projects constantly. When I started using Claude Desktop and Claude Code in my workflow, something unexpected happened: markdown files started piling up everywhere. Skills, agent configs, project notes, CLAUDE.md files — scattered across dozens of project directories. Version controlling them was a nightmare, and every time I set up a new project, I had to dig through old folders to find and reconfigure everything.

I needed one place to keep it all together. Meeting notes, code snippets, daily logs, reference manuals — not spread across 15 different repos.

So I tried everything. Notion, Obsidian, Evernote, Bear, SimpleNote, even Coggle for mind maps. They're all genuinely great tools. But either they required a subscription for the features I needed, or they couldn't quite match the workflow I had in mind.

So I thought: why not just build the note app I actually want?

And here's the funny part — after spending months building Novi Note, I finally understood why those other apps couldn't do what I wanted. Turns out, building a note app that fits your exact workflow is... really, really hard. 😅

What makes Novi Note different:

• 🤖 AI Native via MCP — Connect Claude to your notes with minimal setup. I worked hard to make the MCP configuration as painless as possible. Your AI can read, create, and organize your notes directly.

• 💰 One-time purchase — No subscriptions. Ever. That's a promise, not a marketing line.

• 🔒 Local-first — Your data stays on your Mac. No cloud dependency. Full offline support.

• 📋 Built for work — Daily notes, manuals, documents, post-its, calendar — the structure a working professional actually needs without Notion-level complexity.

I built this for people like me — developers who live in the terminal and IDE all day, use Claude as a daily companion, and just want a clean, private place to keep everything organized without fighting their tools.

Would love to hear your thoughts. What's your current note-taking setup? I'm genuinely curious!

Congrats! Building the tool you couldn't find elsewhere is the best reason to ship something tbh.

Curious whether you're planning to expand beyond Claude to other AI assistants, or keeping it Claude-specific by design?

Looks interesting. What's the price?

Local-first AND one-time purchase? Take my money honestly. I've been bouncing between Obsidian and Apple Notes for months and neither fully clicks. The MCP integration with Claude is what got my attention though , being able to just talk to my notes without some janky plugin setup sounds great. Downloading now.

Zero config setup is underrated — most note apps make you spend an hour tweaking before you can actually use them. How does it handle search across large note collections?

Hi Product Hunt! 👋 I’m Eero, the maker of Sanota.

Family stories shouldn't be dusty archives saved only for the future. They should connect us today. There is no better feeling than a loved one sharing their latest story into the family WhatsApp group. Knowing they had the time to reflect on a memory, polish it with Sanota, and share it to spark conversations right now. That’s the magic we wanted to create.

Try it out, give voice to the most important stories, and fall in love with the experience ❤️

P.S. In Finnish, "Sanota" means to give voice to your thoughts.

Can multiple family members contribute to the same story from different locations, or does one person record at a time? Beautiful concept, good luck with the launch!

Storytelling is all what we need! Really congrats on this level of solution. Wish you all the best here Eero!

Congrats on the launch! The idea of publishing stories one by one as they're ready rather than waiting for a finished archive is a genuinely better way to think about this.

The phone call recording option is a thoughtful touch for older family members. Does the AI adapt its interview style depending on the age or comfort level of the person recording?

Lovely idea, congratulations on the launch!

Hello!

I tried to check it out in my husband's family group on whatsapp (they are Syrian, im Mexican), I played some voicenotes, some recordings but when I came back to Caplo the record mentioned zero words! I also tried with youtube and same zero answer.
I tried several ways but the app never came back with any results. Am I doing something wrong? Maybe It would be useful to have a small tutorial at the beginning or some visuals that show what to expect.

Super impressive, congrats on the product and the launch!

What's next on your roadmap? Do you want to focus on covering more languages, add other complementary features, or something else?

Real-time translation for live calls is genuinely hard to get right. How does it handle strong accents or fast speakers?

If I rewatch a video I've already transcribed before, does it use up my minutes again? Or can I access the previous transcription from the session history?

Love that this is built on ImageMagick and FFmpeg — battle-tested tools wrapped in a modern, effortless interface. Media optimization is one of those things every web developer needs but nobody wants to spend time configuring manually. The fact that it's open source and free is a huge plus for the indie dev community. Does it handle batch processing well? I often need to optimize dozens of assets at once when shipping updates for my browser extension.

Does optimo preserve metadata like EXIF data after compression, or does it strip everything by default? Nice open source tool, congrats on the launch!

Always happy to see an education-related app launching on PH! Congrats - onwards and upwards 🔥

I like the approach of having all the content reviewed by subject-matter experts. I'm curious, for niche exams with smaller pools, how do you find the right SMEs?

Aaand also curious - have you considered expanding towards peer study groups? Learning can feel isolating, and it could be an interesting extension given your focus on personalized content.

Prepping for a computer architecture exam right now, and drowning in course slides makes me realize how desperately the education space needs this. A tool that curates that kind of chaos into a personalized study plan with instant explanations is exactly what's missing when studying solo late at night. Huge congrats on the launch, Tudor!

@tudor_bularca Really interesting approach, especially solving for smaller/niche exams where traditional prep doesn’t work.

One thought — since you're covering 10,000+ exams, do you think new users might need a bit more “proof” or clarity when they first land?

Like showing a very specific example (e.g. how Educato prepares someone for one exact exam step by step) could make the value click instantly.

Feels like once that trust is built early, conversions could improve a lot.

Congrats on the launch and on the mission!

Making quality exam prep accessible to students who may be priced out or simply overlooked, in my view, is genuinely important work.

Would you ever consider a 'study buddy' mode where two people prepping for the same exam can track progress together?

Do you plan to let users import their own question sets for niche exams that might not be covered yet? Looks great, good luck with the launch!

How does the C1 handle thermal management at sustained workloads given the 75W envelope and fanless design? Really cool hardware, good luck!

Can EEVEE track charging costs separately for home vs public stations so you can compare cost per kWh easily? Congrats on the launch!

DailyTips.dev is one of those simple ideas that quietly solves a real problem: staying sharp as a developer without burning hours scrolling docs or tutorials.

Great work and good luck!

As a bootstrapping product maker I can't recommend this newsletter enough. Great work team. Good luck with the launch

I get emails from this newsletter and each single on I end up reading, one of the best newsletters to subscribe to.

Been subscribed for a while and love the daily insights 🔥

Congrats on the launch guys!!

Hands down one of the most practical newsletters for founders. I read every single issue from top to bottom.

Are the tips based on your own experiences shipping products or do you also interview other founders for real world examples? Looks like a great resource!

How does the Bix Match AI handle cases where a buyer's request is vague or spans multiple product categories? Congrats on the rebuild!

What criteria does the AI use to decide it should escalate to a human instead of continuing the conversation? Excited to try this out!

Does it support custom date range comparisons, like this week vs same week last month, all in that single screen view? Clean idea, good luck with the launch!

Hey PH! 👋

GA4 broke something that used to be simple.

You'd open Analytics, see your numbers, feel good (or bad), close it. Done. Now you need a PhD in data exploration just to find out how many people visited your site yesterday.

I got tired of it. So I built better-ga.

Connect your existing GA4 accounts - takes 30 seconds, no new tracking script, your historical data stays intact. What you get on the other side is a single screen that actually makes sense. Users, sessions, pageviews, top pages, top sources, real-time visitors. All the stuff you were opening GA4 to find anyway, just... right there.

The part I'm most excited about: if you're running multiple sites (and let's be honest, you probably are), you can see all of them in one place. Switch between accounts/properties in one click. No more logging in and out, no more 14 open tabs.

It's free to start. I'd love to hear what you think, especially what I should build next.