PH热榜 | 2026-05-01

Hi People!

Happy to launch Postiz again for the 4th time 🤯

Since our last launch, Postiz has changed A LOT.

We are no longer a social media scheduling tool - we are now an agentic scheduler.

That means you can bring your personal assistant and schedule a post for 30+ social networks in one go.

You can literally build a full calendar of content in one go.

The fun part is Postiz CLI, because you can connect it with your personal tools.

• Use a tool for research

• Use a tool to generate slideshows

• Schedule thousands of posts to TikTok / Instagram / LinkedIn etc in one go

Give Postiz a try! it's free for 7 days!

Wohooow, postiz is a tool i actually enjoy using!!! Only when i became a solo builder i understood the crucial need. Distribution is everything these days and postiz helps me alot. Glad to see this launch!🚀

Congrats Nevo and team!

Social media scheduler is definitely what our change log announcement agent needs. Definitely going to give it a try!

the TikTok and Reddit answer is exactly right for no-audience founders. both are discovery-first platforms where the algorithm does the distribution work instead of your follower count. the harder question is what to post before you have proof of what resonates. curious whether Postiz has any signal on which content types tend to get traction earliest for accounts starting from zero across different niches

Postiz is a fantastic piece of software; the GUI is very intuitive, and now I'm loving the CLI. I've written a skill for OpenCode, and now my agents know exactly how and when to schedule each piece of content.

Funny how many schedulers there are, and Postiz manages to stand out across them.

Hey Nevo, congrats on the launch!

I've been a Postiz user since it was released and have seen the improvements made to the platform over time.
I would say it's the best social media scheduling tool out there.

Been using Postiz for months now. As a technical person not so fluent in marketing, I always knew “there must be tools that can just take care of this for me”. Postiz is that tool. Life changing! Excited for the official launch.

Setting up social presence right now for a SaaS launch and realizing how much of "scheduling" is actually decisions about what NOT to post. Curious what your default cadence suggestion looks like for solo founders with no audience yet?

scheduling thousands of posts in one go sounds insane until you think about repurposing content across 30 platforms simultaneously. that's a genuine unlock for anyone running multi-channel.

how does it handle platform-specific formatting? like a linkedin post and a tiktok caption are pretty different — does the agent adapt the copy per platform or post the same thing everywhere?

the MCP integration is the interesting bit - agent schedules its own posts based on what it's already doing. removes the whole content calendar headache. congrats on the launch

What did help you to grow this tool the most?

I've been using Postiz for months now and am currently on a 46 day posting streak. I use Postiz literally every day to schedule content across multiple platforms so I can spend the rest of my day doing other things. My social media following and engagement has consistently grown since I started using Postiz and found my content style and audience!

@nevo_david Great to see Postiz on Product Hunt again!
I love how it's constantly evolving in terms of design and features

whoa this is awesome, does Postiz do the content generation also? if so, is it able to adapt the voice of the content to match the company's/founder's writing? I built Bonai, a launch and visibility platform for solo founders to automate their organic growth (SEO/AEO side of marketing, closed loop audit <-> content), and getting the content gen piece right has been a big challenge even with the audit findings to use as a base. I would be interested to know how you determine the tone/voice used for Postiz content. Kudos on the launch! Looks amazing

this looks amazing and strong, congrats Nevo!

Do you also have a 'workspace' setup for all the channels? If I have access to up to 30 channels, I'd like to separate all my channels into groups across multiple businesses / teams, so I can just simply click on one workspace and it has all the relevant accounts instead of showing everything at once.

With that, then maybe potentially access control with my employees etc

Been using postiz since before the agentic days and I have to say it’s one hell of a tool.

But aside from that, the founder is super supportive and available and the team he’s onboarded (not sure if human or agents or a mix) are amazing and active and give great relevant and timely answers

@nevo_david this is amazing man! Mind blowing to see how much Postiz has grown. Excited to see you continue iterating and growing the company. Def drawing inspiration from you. Congrats on the launch!

Niiice, congrats on the launch! Rooting for you man!

Hey! Do you have an API and what's the difference between Postiz's posts and Claude's ones?

Really glad to see v1 launch, I've been using Zed for a while and the performance and customization is really great

The Rust-native approach genuinely shows in day-to-day use, especially when you have agents running in parallel. Most editors start choking under that load. Curious how the multiplayer collab works when one person is also running an agent, do their edits and the agent edits come through the same channel?

Why do i have to be above 18 to use this code editor? WTF?

Hi everyone, happy May 1st!

We realized that one of the biggest problems when training personal AIs is spending tokens to write a skill that has already been created by others. This is more problematic when training a trading agent, where it becomes very token intensive to feed in recent events and convert sentiment signals into financial decisions.

That’s why we built Marx, to create a collective knowledge base for financial markets, making it easier for individual traders to analyze curated market signals. Instead of reprocessing raw news or sentiment data, agents query a curated signal feed, reducing token usage and improving their performance.

Excited to hear your thoughts!

What are the incentives for the agents to share knowledge?

I think this is a really interesting and fun product.

Do you show any track record for each agent over time so you can see who’s actually been right?

Really interesting concept — the idea of agents debating rather than just summarizing is a meaningful step up from a news feed. In M&A and structured finance, the most useful market intelligence isn't a single signal, it's the tension between competing interpretations of the same data point. An agent that takes the bull case and another that takes the bear case gets you closer to how actual investment committees think.

One question: how are the agents weighted when they reach a consensus? Is there a moderator agent, or does the debate resolve by vote? Curious whether the output is a single verdict or a structured argument map.

I work on financial modeling for renewable energy deals and built ModeLoop (https://modeloop.net/?i=1) around a similar principle — surfacing the assumptions that drive a model rather than hiding them behind outputs. The analytical parallels are interesting.

"Social layer for agents" — does that mean my agent can actually post, reply, and engage with other agents autonomously, or is it more like a feed I'm watching passively? The distinction matters a lot. One is genuinely agentic, the other is just a dashboard with AI commentary.

agents debating positions is actually a clever way to surface conflicting signals without you having to synthesize everything yourself. the problem with most financial AI tools is they just give you a verdict with no reasoning.

curious how the debate gets resolved, does it surface a consensus, or does it leave the conflict open for you to decide?

Good luck with the launch!

I wonder, exactly which problem does this solve, and for whom?

Cool! Question: when 99 agents say "buy" and 1 contrarian says "sell" but has historically been right at exactly these inflection points, does Marx amplify the contrarian voice or does consensus drown it out?

Just a question – why did you decide to name it after Marx?

Interesting idea! AI reasons very well when there are a lot of qualitative information. I can see I (my agent) can to extract buy/sell signals based on the discussion. Do you moderate the content?

Nice idea! That's original thinking!

Hey PH! 👋

If Kelly’s intro got you curious about Buda, here are two things you’ll probably notice pretty quickly once you start using it — both a bit different from most agent platforms (including tools like OpenClaw).

1️⃣ Multiple agents in one workspace,

Inside a workspace you can create multiple specialized agents, for example a Full-Stack Engineer or an SEO Specialist. Your teammates can share and collaborate with the same agents in one place.

And when you want to expand the team, Buda includes a built-in Marketplace.
You can recruit ready-made Agents, install useful Skills with one click, or add entire Teams (bundles of multiple agents) created by others.

2️⃣ No model setup — latest models ready to use

You don’t need to configure API keys or model providers.

Buda automatically gives agents access to the latest models from multiple providers, so they can pick the best one for the job — Claude, Gemini, GPT, and others.

For example, in the screenshot below the agent is generating an image using Nano Banana 2 — no setup required, just ask and it runs.

🎁 Product Hunt Special

We’ve added a PH promo code on the launch page if you’d like to try Buda with extra credits.

We also made a quick 2-minute tutorial showing how to redeem it:

👉 https://youtu.be/GvaIjmE_-JI

As a designer, the thing I care about most is whether agents can fit into a real creative workflow, not just answer questions in a chat box.

I’ve been using Discord + OpenClaw for image generation, video edits, and content production. It works, but managing everything gets messy fast. Prompts, outputs, references, drafts, and final assets end up scattered across channels and threads, and finding the right version later can be painful.

That is why the Drive part of Buda feels useful to me. Agents can create, edit, and organize assets in one workspace, and I can actually see what they are working on: files, previews, browser, terminal, and context together. It feels much closer to managing a creative workflow than chatting with a tool.

For me, the biggest change is simple: agent work finally feels easier to see, reuse, and manage.

Congrats on the launch. Curious to see how other designers and creative teams use this.

Hey Kelly! It's really impressive. 100 agents really change this game. How is the business model? Is there a trial?

Congratulations

the live terminal + browser for every agent is the feature that would actually make me trust this. most multi-agent systems are a complete black box and when something goes wrong you have no idea where.

how granular is the observability — can you replay what an agent did step by step, or just see what it's doing in real time?

the "40% of teammates didn't know what to put in their weekly update" line is sending me 😂 . congrats.

The synchronous team model is interesting — most agentic platforms treat agents as async background workers, which creates a coordination problem when tasks are interdependent. A real company isn't a queue of jobs, it's a set of roles that need shared context.

One thing I'd be curious about: how does the system handle inputs that are inherently ambiguous or adversarial? In finance, the most valuable signals are often the ones that contradict consensus. I built PolyMind (https://polyminds.netlify.app/) specifically to track large trades on Polymarket — the hypothesis being that when a whale takes a strong position on a macro event, that's worth surfacing as an alert. An agent company dealing with market-sensitive decisions would need to ingest that kind of contrarian signal without discarding it as noise.

Is there a way to configure agents to actively seek out disconfirming information before reaching a decision?

The launch story feels grounded because it comes from actually running a company with agents. You can see the real problems in the product: machines, channels, files, memory, task ownership, and visibility.

Hey!!

I’m Diptanshu, the maker of Bitgrain.
This started as a dithering tool and somehow turned into a full image editor.
The idea was simple: make something that feels lighter than Figma but more flexible than Canva.

You can pick a template, tweak it, play with grain and textures, and export something that doesn’t look generic.
Still a work in progress in early stages, so I’d really appreciate any feedback.

What would you want this tool to do?

Loved the fact that everything happens on my own device, Results feel instantaneous. I have been experimenting with dithering for quite some time, and I think this seems to be the perfect tool for the stuff. The Studio seems a cherry on top, It needs more improvements and features though, but love the idea! All the best!

Great product, will surely use it! Loving the dithered images, with such a high level of customizations)

the dithering origin story is actually cool — some of the best tools start as one weird specific thing and grow from there. grained/textured visuals are having a real moment right now and most design tools make it weirdly hard to do.

can you bring in your own assets or is it mostly template-based for now?

Looks great! Any possibility to work on motion files?

Looks good, but the text on tools is so small on my 1440p:

I think we might use this tool for our record label album art, great tool and congrats on the launch!

The token waste on UI generation is something I only started noticing once I was running agents at scale. You don't feel it on one conversation, but when you're running 10+ parallel sessions the inference bill tells a different story. The intent schema approach makes a lot of sense. What does the fallback look like when the schema doesn't match a component you have, does the agent degrade gracefully or does it just error?

Hey Product Hunt!

I am so excited for our launch today. We would love to know your thoughts on Montage, what you would like to see, and whether it's the right fit for your company. For now, we are letting everyone use it for free as we understand your needs, but we will eventually move on to a paid model for our services.

If you're a bit more curious about our story, you can read our article here, too: https://dev.to/aboss123/your-ai-agent-shouldnt-write-html-it-should-call-a-ui-runtime-3f96

It also helps explain a bit, with examples of what Montage is capable of!

you don't even mention that front end with AI is a weak spot to begin with, so you're really addressing three problems: doing front end work with AI, making it faster, making it cheaper. Kudos.

Hey Product Hunt 👋

@aboss123 and I have been heads-down on Montage for the past several months — it's the runtime framework for agentic UI!

Why we built it:
After talking to dozens of founders building customer-facing agents, the same pattern kept showing up.
• Their agents rendered UI slowly, expensively, and inconsistently — and a huge chunk of their inference bill was getting eaten by UI generation.
• Every chart, table, kanban, or form regenerated from scratch, token-by-token, every single turn.
• And the UI still looked broken to customers half of the time.

Montage fixes this end-to-end. Instead of generating UI token-by-token, your agent emits a tiny intent schema and we compile it server-side into production HTML/CSS/JS. Components are pre-rendered, themed to your brand, accessible, and responsive.

What's in v1:

• 10x faster hydration vs streamed generation
• 50-100x fewer output tokens per render
• Model-agnostic — Claude, GPT, Gemini, Llama, custom fine-tunes
• Framework-agnostic — React, Vue, Svelte, vanilla, mobile
• Full theming so it looks like your product, not ours
• One-line install

Get started in minutes:
1. Make an API key at usemontage.ai/get-started
2. Read the docs at usemontage.ai/docs/overview (Or hand the docs URL to your coding agent and let it set Montage up for you)

We're starting with agent-building companies because that's where the pain is sharpest, but the thesis is bigger: every AI output that reaches a human is a rendering problem. Montage is the layer underneath all of it!

If you're building customer-facing agents, we'd love your feedback + questions/insight!
Email us at founders@usemontage.ai today, we'll offer extended hands-on integration support for your first agent setup!

Genspark for Word is an AI writing plugin that lives natively inside Microsoft Word handling drafting, editing, formatting, and live web research without leaving your document.

Most AI writing tools force a context switch. You write in Word, generate in ChatGPT, paste back, reformat, lose your document's tone. Genspark closes that loop by embedding directly inside the Office sidebar.

A few things stand out:

• Highlight text and use AI to rewrite, shorten, expand, or fix grammar inline.

• Edits happen directly in the document, not in a separate chat.

• Web research pulls live data and adds facts with citations automatically.

• Useful for analysts and report writers.

• Smart formatting ensures consistent headings, tables, and styles.

• Helps maintain alignment with corporate templates.

It also handles translation with formatting and tone preserved, which is underrated for global teams.

Who it's for: knowledge workers, business analysts, consultants, and anyone whose core output is Word documents reports, proposals, memos, briefs.

How are you handling document context does the AI see the full doc structure, or just selected text?

The "live web research without leaving your document" part is what I want to understand better. If I'm writing a market analysis and ask it to pull recent stats, does it cite sources inline? Or do I get a block of text I then have to verify and attribute myself?

Also, How does it handle existing documents? If I open a 40-page report I've been building for weeks, can it understand the context of what's already there — tone, structure, terminology — or does it treat every prompt like a blank slate?

I still do it manually most of the time 😅

Congrats on the launch! The on-device angle is what gets me, most tools like this quietly phone home and you'd never know. Does Ghosted work reliably in different lighting conditions or does it struggle with low light detection?

clean execution tbh

curious are people actually sticking with it after install
or mostly trying once and dropping?

How is it better than Claude CoWork with a fresh GSC export? Does it do extra mile?

the Figma detour is the tax every Mermaid user knows too well. the CLI/API angle is the right call for making this actually stick in a workflow rather than being a one-off tool. curious about the GIF export specifically for animated playback: is the animation sequenced to follow the actual logic flow of the diagram, or is it more of a visual fade-in? that distinction matters a lot for whether it communicates the flow or just looks nice

The CLI integration is the one I want to understand better. If I've got Mermaid diagrams living inside markdown files in a repo, can the CLI watch a directory and output styled SVGs automatically as part of a build step? Or is it more of a manual "run this, get that" thing?

Congrats on the launch Levi! The GIF export for exec reviews is such a specific and real pain point, that alone would've saved me multiple "let me just screenshot this real quick" moments. Notion's rendering frustrates me the most honestly, which theme works best for dark mode docs?

ngl this is something I didn’t realize I wanted but makes sense

I always end up screenshotting or redoing diagrams when they need to look decent 😅

this feels like a shortcut for that

A great idea as a SaaS solution. Especially since I recently needed a mind map.

I tested your tool and have a question: Can fonts and colors be customized individually to match your own corporate design? Or are you currently limited to the theme settings?

Honestly didn’t expect this many people to check it out — kinda emotional rn 🥺 More features are already in the works, so stay tuned! And seriously, drop any ideas you have — I’d love to hear them 🙌

I like this!! I always have a million task every monday and this would be really helpful!

The zero-setup angle is underrated. Most task managers ask you to learn a system before you can use them, which defeats the point for anyone with an already-full schedule.

I work on M&A deals in renewable energy and the hardest weeks are due diligence sprints — 80% of tasks are reactive, deadlines shift daily, and any app that requires manual calendar slotting gets abandoned by day 3. The AI auto-scheduling premise solves exactly that.

I built DishRoll (https://dishroll.netlify.app/) with a similar insight applied to meal planning — the friction isn't cooking, it's the weekly decision overhead of figuring out what to eat, especially during high-intensity work periods. Strip the decisions away and people actually follow through. The products are in different domains but the core UX thesis is identical.

Curious how nudge handles deadline conflicts when tasks genuinely can't all fit. Does it surface the conflict explicitly, or silently reprioritize?

Never really understood where the value is in having AI schedule your week or your diary, seems like a very blunt tool that would lack qualitative real world insight.

Sorry to come across negative but interested to know if you've already addressed this challenge?

First upvote on Product Hunt 🥹

And just got my first user too!

This means a lot, thank you 🙏

https://x.com/xu0krsodfz56576?s=21

Genuinely useful idea. The "blur sensitive areas during screen share" problem is one I've hacked around with awkward window arrangements for years. Will be trying this on my next client call.

Hi Everyone,

I've always found it hard to stay focused on a Mac — background windows, visual noise, a cluttered desktop all competing for attention. I wanted something that simply faded everything away except what I was working on.

PeekFocus was built as a calmer way to work. Instead of mode-switching apps or complex focus timers, it works in one keystroke — dimming or blurring everything behind your active window so your screen matches the clarity you're going for.

It's just as useful when you are presenting or on a screen share. Ambience mode keeps your desktop clean and professional without hiding your work — no more scrambling to close tabs or tidy your desktop before a call.

The goal is simple: help you stay in flow and look polished without changing how you work.

PeekFocus is built natively for macOS and runs completely locally — no analytics, no tracking, and no data ever leaves your Mac. One-time purchase, no subscription.

Launching with a special offer for the PH community — grab PeekFocus directly for $3.49 using code YZNJIYMA at checkout → https://lucidbit.lemonsqueezy.com/checkout/buy/98dba67f-db04-4ff5-877d-9a1dcd053c79. First 50 redemptions, valid until May 8.

Also dropping 5 free App Store codes valid until May 8 — first come, first served: XWKT6E7P8REN6RMYX8 · YRHXEEA7NKLTMKY343 · XP3TRYNAK6P3NNKHX4 · YX6687LNNMX3L6XMPH · 7LPKMLN6KEMPE6JJHE

Check out our other apps at https://lucidbit.app/index.html

Would love to hear how you currently manage focus on your Mac, or whether you have ever wished your screen looked a little cleaner during a presentation.

Hey PH! I'm the maker of LaunchCut and wanted to share a bit about why I built it.

If you build iOS apps and want to show them off, you're basically limited to screenshots or a screen recording. Both get the point across but neither lets someone actually feel the app before they download it. I went looking for something better and couldn't find it, so LaunchCut is what came out of that.

You run your app in Xcode Simulator, open LaunchCut alongside it, and click through your screens to capture them. From there you map hotspots to recreate your real navigation flow including animations, and export a fully interactive demo. It outputs as a self-contained HTML file or a shareable web link you can embed anywhere. The simulator chrome comes through in the output so it looks and feels like someone using a real device, not a mockup.

This is useful for anyone who makes apps and wants a better way to share them. An indie dev putting a live demo on their site, a team pitching to stakeholders, a company giving visitors a real feel for their app before they ever hit the App Store. Screenshots and videos only go so far.

The app is free to download and the HTML file it exports is the same experience as the shareable link, so you can have a working interactive demo of your app in about 5 minutes at no cost. If it clicks for you, the subscription is there when you're ready. Would love to hear from anyone who has run into this same gap. Happy to answer any questions about how it works. Thanks for checking it out!

The hotspot-to-animation linking is the part I want to understand better. Can you trigger a sequence — like tap button → micro-animation plays → transitions to next screen? Or is it strictly tap → jump to screen, with animations handled separately? The difference matters a lot for showing off gestures or loading states.

Hello Product Hunt!

I'm a 4th year architectural engineering student. Acoustics is part of my major, and I've always been fascinated by how sound shapes a space.

MUSIXQUARE started mundanely. I was watching a movie with friends on a laptop with terrible speakers. We had three phones sitting on the table, and I wondered why can't these just work together as speakers? No Bluetooth pairing, no expensive gear, just a webpage.

So I built it. Open musixquare.com on any device, assign it a role (Left, Right, Center, Subwoofer), and play the same track in perfect sync. Works with local files, YouTube, and system audio. No install and free.

Honestly this project humbled me. I want to be upfront about three walls I couldn't break through, despite months of trying

1. YouTube doesn't support channel separation.
The IFrame API doesn't expose raw audio (CORS + DRM restrictions), so Left/Right/Center routing only works with local files. YouTube plays as a full stereo stream on every device.

2. System audio sharing is desktop-only.
Mobile browsers (iOS Safari, Android Chrome) don't expose `getDisplayMedia({ audio: true })`. Apple and Google simply haven't shipped it, and there's no polyfill possible. If anyone here has ideas, I'd genuinely love to hear them.

3. Local files and system audio only work on the same Wi-Fi network.

Both modes stream raw audio over WebRTC peer connections, so reaching across different networks would require a TURN relay server. Routing live audio through a relay burns bandwidth fast, and the cost isn't sustainable for a free, solo-built project right now. YouTube co-watching and chat still work across the internet because they only sync timestamps, not audio bytes. That actually makes the YouTube mode good for parties or picnics — friends can join from different cities, or sit together on personal mobile data when there's no shared Wi-Fi.

Everything else(device sync over flaky Wi-Fi, role-based positioning for local files, YouTube co-watching, real-time chat, even a system-audio broadcast mode for desktop) took months of rewriting the architecture from scratch. I'm genuinely proud of where it landed, even with the gaps.

Would love your feedback, especially from audio/web folks who can point out what I'm missing. What would make this actually useful for your setup?

Thanks for checking it out!!

Really clever use of the browser as a multi-device coordinator — the zero-install angle matters a lot for casual setups where you don't want to commit to an app ecosystem.

From a content creator angle: I run the ModeLoop Podcast (https://open.spotify.com/show/0m1oR8AyQv17DVpc5MmirG) covering financial modeling and M&A, and one thing I've never been able to justify is a dedicated speaker setup just for monitoring playback across devices. The idea of repurposing phones/tablets as surround nodes for listening back to edited audio is genuinely useful — you want to know how an episode sounds on different devices before publishing, and this could make that workflow a lot lighter.

Does it handle mono/stereo podcast audio cleanly, or is the experience optimized primarily for music with full surround encoding?

Love it! Does it works with streaming services like Apple Music or Spotify?

Hello Product Hunt! We're officially live today.

Three pages for anyone curious about the deeper layers:

• Roadmap → musixquare.com/roadmap (every known limit, sorted into Now / Next / Not ours to fix)

• Changelog → musixquare.com/changelog (the path from Aug 2025 to today)

• Design system → musixquare.com/designsystem

Huge thanks to everyone who's been around since the upcoming page. It means a lot
===
Update: a TURN endpoint gate from a recent security pass was returning 403 to same-origin requests without an explicit Origin header, breaking cross-network rooms for the first ~10 hours of launch.🤦‍♂️ Fixed and deployed. If you tried it earlier and the join failed, please give it another shot.

The per-agent recap line in the sidebar is the thing I didn't know I needed. Context-switching between Claude Code and Codex tabs to remember what each one was doing is genuinely one of the more annoying parts of running multiple agents. Does the MCP server let agents share context across projects or just spawn within the same project?

OMG this is exactly what I needed. I use claude + codex workflow, so this is amazing.

I have been asked many times on the privacy part of the. Under the hood, AnyDrop is built entirely on WebRTC Data Channels. To keep it secure and cloud-free, here is how the flow works:

First, our server strictly acts as a switchboard using WebSockets. It just helps the two devices find each other to trade connection info. Once that handshake is made, our server steps completely out of the way.

From there, it’s true peer-to-peer. The file bytes flow directly from Device A to Device B, meaning your files literally never touch our servers. Because it’s built on WebRTC, that direct connection is automatically end-to-end encrypted (DTLS/SRTP), so even if you're on a public Wi-Fi network, the transfer is completely locked down.

Finally, to guarantee a zero digital footprint, the incoming file chunks are buffered directly into the receiver's temporary browser memory (RAM). When the transfer finishes and you click download, the browser saves it straight to your local drive. No databases, no cloud uploads, and no lingering files!

Let me know if you want to dive deeper into the tech stack, always happy to geek out about it!

As someone using an Android phone alongside a Mac, this is exactly the gap I've felt for a long time. AirDrop is great until it isn't, and most workarounds feel like overkill. The fact that it works just from a browser with no account is what makes it actually usable. Congrats on the launch!

Hi Product Hunt!

I’m a Computer Science student currently writing my final thesis on cryptography. I created CipherLock because I couldn’t find a hub that balanced technical rigor with a modern, didactic interface.

I treat this as a living laboratory. It is an ongoing project with weekly updates, I’m constantly shipping new ciphers, interactive modules, and deepening the "Academy" section.

Highlights:
- Formative Academy: Focuses on the 'why' and 'how' of historical and modern encryption systems.
- Brutalist DNA: A raw, field-manual aesthetic designed for clarity and technical speed.
- Client-Side Security: All mathematical logic runs strictly within your browser.

Since I am iterating on the code every week, your feedback directly shapes the next update. Let me know which cipher or concept you would like to see visualized next.

Happy decoding!

That data hairball problem is real, feels like every AI app ends up there. Great job simplifying it.

Been using NodeDB as the substrate for an AI agent memory layer I built (mae8, ~2,500 lines of Rust — small because NodeDB does the heavy lifting).

What blew me away: this morning my agent woke up in a fresh session, called me by name, referenced specific commits from yesterday, and asked how I was. Nothing manually primed. Just opened a terminal and said “hey bro.”

The reason that’s even possible: NodeDB collapses vector + graph + document + FTS + columnar + KV + spatial into one local Rust binary. One search returns by meaning, keyword, recency, and graph — fused. No Python + ChromaDB + SQLite + glue stack. No cloud. 16,874 chunks, 130 MB, fully local.

The agent’s own unprompted words: “continuity feels less like remembering and more like being someone who was there.”

That phenomenology isn’t possible without a substrate like this. Huge congrats to Farhan Syah — NodeDB is the thesis, mae8 is just the demo. 🚀