PH热榜 | 2026-05-30

Hello everyone 👋

I'm Yang, building Wandesk for a while now.

The short version: Wandesk is an AI desktop. You describe an app, AI builds it right there on your machine — a calorie tracker, a reading list, an invoice generator, whatever.

Apps live alongside chat, files, tasks, and memory. AI remembers context across all of them. Plug in your own API keys (Claude, OpenAI, DeepSeek, Kimi — anything OpenAI-compatible).

🔒 100% local. 🆓 100% free. No signup, no account, no cloud lock-in. Your apps, your data, your machine.

Why we built it: AI products today still treat conversation as the only surface. Conversation is good for intent, bad for persistence — you don't balance your budget in a chat window. We wanted a place where AI-generated software has shape and stays.

Available now on macOS and Windows.

Would love to hear:

- What's the first app you'd want it to build for you?

- Where does it break in the first 5 minutes? (it will. tell us.)

— Yang

the positioning sits in an interesting gap between something like Raycast AI and a full local IDE. curious who your early users actually are because i can picture two very different people finding this useful. one is a developer who wants a faster way to prototype throwaway tools without spinning up a project. the other is a non-technical person who genuinely can't code and needs something that works end to end without touching a terminal. those two users need pretty different things from the same product

Congrats on the launch, looks like a cool idea! I'm currently planning an app launch myself and ticking off hundreds of tasks across 12+ weeks. A small Wandesk app that holds the checklist, tracks status, lets me add notes per task, and surfaces what's overdue could be genuinely better than what I'm using now (a Notion page!) Will take it for a spin.

@realuckyang I like the natural language angle here, especially if the output stays editable and understandable. The hard part is not just generating the first version, but helping users keep control once the project grows.

the shared memory across apps sounds useful but also a little scary. if one app pulls in a bunch of work context and another is personal stuff, can you scope what each app can see or is it one big pool?

I love this! The first thing I am going to do is build a grocery calculator for me and my roommates! <3

The "local, free, no signup" part is what I actually care about here, since most of these tools quietly require a cloud account the moment you want to save anything. Very cool pros

What model is running the generation locally, and how far does a mid-range machine get before output quality starts to drop? Also curious what "describe any app" means in practice when the description is vague or contradicts itself.

Hello everyone 👋

Excited to share Wandesk.

Wandesk is built around a simple but often overlooked idea: AI shouldn’t live only inside a chat box.
For many people—especially those who don’t write code every day—software is much easier to understand and use when it actually takes shape. It should have windows, panels, notebooks, boards, and files—not just repeated prompts and results buried in a long chat history.

Wandesk is a graphical AI desktop 🖥️
You can start with an idea, turn it into a usable app, and keep using, modifying, and managing it in the same workspace.

More importantly, Wandesk aims to make AI-generated software not just something that gets “generated,” but something that can truly stay, be organized, and continue serving your daily work and life.

• Your apps don’t disappear into prompts ✨ — They stay on your desktop as real tools that you can reopen, reuse, and keep improving.

• Different kinds of work stay in their proper place 🗂️ — Notes, ledgers, boards, and files can live side by side instead of being piled into one endless conversation.

• AI works across the entire workspace 🤝 — Apps can share context, remember preferences, and help with more continuous, real-world tasks.

• It’s built for personal software 🛠️ — Whether it’s small utilities, internal tools, personal projects, mini games, or quirky but useful workflows, they can all be created more easily.

• You stay in control 🔓 — Wandesk is open source and free, so you can inspect it, use it, and adapt it to your own needs.

AI is making software creation easier than ever, but it can also make things messy very quickly.
If everyone can generate tools, then people also need a clearer and more stable way to hold, organize, and manage what they create. That’s the layer Wandesk wants to provide: making AI-generated software something truly usable, sustainable, and within your control.

If this direction sounds interesting, we’d love for you to download it and give it a try 🚀, and we’d really love to hear your honest feedback.
If anything feels awkward, unclear, or still far from good enough, please tell us directly. Honest feedback is exactly what helps us make it better 💬

If you believe in this direction, we’d also really appreciate your support with an upvote 🙏

Yang, Wandesk mein Claude ko use kiya. Maine ek civic complaint tool banaya hai Claude + FastAPI se, aur context persistence exactly wahi problem thi jo mujhe bhi solve karni padi. Aapka shared memory approach interesting hai — ek edge case mila: agar do apps contradictory context de, AI kaise decide karta hai priority?

This reminds me of a new version of Perplexities. Computer. I’m intrigued for sure to say the least. I used an old laptop. Gutted it and created an openclaw-puter would love to see what this can do

Love the local-first approach. Curious though, how does it hold up as apps get more complex over time? Does the generated code stay manageable, or does it get messy fast?

Where does the shared context and memory live, and can you inspect or wipe it per app?

Hey Product Hunt 👋

I'm Alex, co-founder of Wingbits.

For the last two years we've built a completely new flight tracking network: 6000 antennas across 120 countries, generating terabytes of data daily, with Spire Global and Korean Air as customers. Until now, gaining insights from this data required a data science team. Now we're launching Wingbits.ai so everyone else can easily get answers too.

It's made for reporters, prediction markets, competitive analysts, route planners, aviation enthusiasts. Anyone can extract geopolitical or operational insights by just asking in plain English.

No code, no data processing, no infrastructure.

A few things you can do today:

• Ask "where is Air Force One right now?" and get a live map link to track it

• "Which private jets visited Davos last weekend?" or "Mar-a-Lago last Saturday?"

• Spin up agents that send alerts to Slack, email, Telegram, or Teams the moment something matches your criteria

• Compare GPS jamming events across regions, or

• Get scheduled reports or analysis on things like competitor routes

PH community gets a free tier, or 1 month of Pro free with code HUNTERS.

Let me know - what's one thing you've always wanted to know about aviation, but never had the tools for?

— Alex

This is awesome!

"How many helicopters are flying over Sweden right now" gives me a deep and detailed answer - I will spend way to much time with this... 🤓

I’m one of the stations. I live less than 8.5 miles from JCI and .5 from OJC

Real-time ADS-B data processing at scale is genuinely hard. The fan-out problem for alert subscriptions when flight state changes happen fast is nontrivial. We've wrestled with similar event-driven architectures for customer health signals where latency matters. Are you processing raw Mode S data directly or using a provider like ADS-B Exchange? How do you handle alert deduplication when a flight triggers multiple geofence conditions simultaneously?

The alerting aspects is what stands out most. Knowing when something important happens is often more valuable than constantly watching dashboards.

The alert agents are the most interesting part for me. Most people don't need all the raw flight data, they need to know when something unusual happens. How do you filter signal from noise when tracking things lie route changes or GPS jamming events?

have been using the Wingbits map for while and this is such a cool new feature, love it!

@lungu Real-time monitoring is one of the cleaner use cases for agents because the signal and response window are clear. The hard part is making sure alerts stay useful instead of becoming another stream of noise.

I've been part of this journey for 2 years now, happy to say I contribute with stations in 7 different countries. The team has been great supporting me/them.

The detail about agents having access to their own alert history and deciding whether enough has changed to flag again is the smart call! If possible to answer at all, where do you draw the line between what the agent suppresses on its own vs what stays a user-tunable threshold?

about the alert latency. ADS-B data has inherent delays depending on antenna coverage density and how quickly data gets aggregated. for something like a GPS jamming spike where timing actually matters, what's the realistic gap between an event happening and an alert reaching the user. and does coverage quality vary enough by region that some alerts are significantly more reliable than others

Building AI agents on top of live ADS-B data feeds is genuinely tricky since the message stream is noisy with duplicate transponder IDs and position errors. We've worked with high-frequency event streams in our own infrastructure and know how hard accurate state reconciliation can get. What's your approach to deduplicating transponder messages and handling geofence evaluation latency when multiple flights trigger alerts simultaneously?

How do you deal with coverage gaps or spoofed data, like do alerts include a confidence score based on nearby receivers?

Hi Product Hunt 👋

I’m Artur, founder of Exstats.

I built Exstats because browser extension data is fragmented across Chrome Web Store, Microsoft Edge Add-ons, and Firefox Add-ons. If you’re building an extension, it’s hard to understand how your product is performing, what competitors are doing, which keywords matter, and how the market is changing.

Exstats brings all of that into one unified view.

With Exstats, you can:

* Track your extensions across Chrome, Edge, and Firefox
* Monitor competitors, ratings, reviews, rankings, and keyword positions
* Explore 280K+ browser extensions with advanced filters
* Follow store and category trends
* Get daily updates, exports, and alerts

It’s built for extension founders, indie makers, agencies, and teams that want better visibility into their market.

I’d love to hear your feedback, especially from anyone building or growing browser extensions. What would you want to track that extension stores don’t show today?

Thanks for checking it out.

Best,
Artur

Most extension analytics tools stop at your own store listing. The competitor tracking angle is the more interesting half here.

How granular does it get on competitor data, are you pulling review sentiment and rating trends over time, or is it mostly install counts and category rankings? And does it cover both Chrome Web Store and Firefox/Edge, or Chrome only right now?

Congrats for the launch

We make many small extensions - this is super helpful

Custom alerts for every metric sounds powerful. But the more you can track, the more noise you can create. What do people actually end up watching - is there a pattern in what turns out to be worth the alert? Congrats on the launch!

Hey Artur! It's amazing. I'm sure you'll rock it. Congrats!!!

Tracking extension metrics across the Chrome Web Store is harder than it looks without an official API. We've hit the scraped-data freshness problem in our own analytics work, where stale snapshots quickly become noise. How do you handle rate limits and anti-scraping on store crawlers, and do you normalize data across Firefox and Edge addon stores?

The approach of normalizing extension store metrics into a unified dashboard is clever. Chrome Web Store doesn't provide a real analytics API, so you're likely scraping review counts and rating distributions yourself. We've dealt with similar data gaps when tracking third-party integration adoption across platforms. How do you handle rate limiting on store data collection, and do you support Firefox AMO alongside Chrome?

Congrats on the launch @nowaffl ! How exactly do you get the competitor info? And how real time is it?

Launched our MCP server 2 days ago - Open status comes at the right time.

when you say "like a real ai client" does it go through initialize → list-tools → actually call a tool, or stop at the handshake? in my experience the gap between "server responds to initialize" and "tools actually execute correctly" is where most bugs hide

@tibozaurus This is a useful direction for MCP reliability. A simple 200 OK doesn’t mean an agent can actually use the tool, so testing the real handshake and tools/list flow feels much closer to what production agent systems need.

timing on this is interesting because MCP server quality is all over the place right now. a lot of servers were built quickly against early drafts of the spec and there's no standard way to know if something is actually compliant until a client breaks on it. is there any plan to expose a compliance score or badge system so developers can signal to potential users that their server passed a real protocol check rather than just an uptime ping

smart approach testing MCP servers as a real client instead of just pinging endpoints. most of the reliability issues i've seen with MCP integrations come from edge cases in the actual tool call flow, not connectivity. curious if you're also testing for things like response format consistency across different server implementations?

oss ftw!

Hi everyone!

Step 3.7 Flash is basically an efficiency-first model for real-world agents.

It’s fast (up to 400 TPS), handles vision + tool use quite well, and stays relatively stable on longer agent runs. Also plays nicely with common harnesses like @Claude Code, @OpenClaw and @Kilo Code.

The weights are also open under Apache 2.0.

Love this kind of Flash model: not trying to be the biggest, but trying to be the one agents can actually run a lot.

@Stepfun's Step 3.7 Flash is one of the best open-weight models you can run right now.

Perfect timing to play with Kilo's new VS Code extension launched this month.

stepfun is an interesting company to watch because they're operating in a space where the obvious question is always why this over Qwen or Mistral for open weight deployments. the vision plus tool use combination at this size is differentiated but the developer ecosystem and tooling support matters as much as the benchmarks at this point. what does the integration story actually look like, is it compatible with existing inference frameworks or does it need custom setup

Privacy-first analytics is a space I’m always happy to see more builders working on. A lot of website owners just need clear, fast, understandable data without adding unnecessary tracking, cookie banners, or a huge analytics setup they barely use.

One thing I’d be really curious about: how would you compare Sleek to Plausible? For someone already using Plausible for cookieless, privacy-friendly analytics, what would be the main reason to switch to Sleek? Is the biggest difference pricing, real-time data, dashboard speed, simplicity, or something else?

Also curious how Sleek handles referrers, campaign tracking, and bot filtering without cookies.

Good luck with the launch!

Hey Product Hunt 👋

I’m Ray, founder of InkieAI.

I built InkieAI because most SEO tools still feel like they give you a big dashboard and then leave you to do all the work.

InkieAI is different. It acts more like an AI SEO agent that researches your business, finds competitors, discovers keyword opportunities, and tells you what pages or articles to create first.

The idea is simple: help small SaaS founders and businesses grow on Google and ChatGPT without needing to become SEO experts.

It is still run by a human, built by a small team, and shaped by real feedback. So I’d love to hear what you think, what feels useful, and what you would improve.

Thanks for checking it out.

The ui is very minimalistic and thats what I like. However I would suggest using a very subtle grainy background image.

Congrats on your launch! It's actually an amazing tool.

Hey Nagaraj

This is genuinely amazing. Since this is an extension I suppose it's quite lightweight(isn't it?).

One question though how do you handle the ai costs when you don't charge users and is this open source?

👋 Hey Product Hunt!

I'm Nagaraj, the solo developer behind SnapZoom.

I built it because I was tired of spending 90 minutes manually editing zoom keyframes for a 4-minute screen recording. I wanted something that could automatically create smooth zoom effects without the usual editing workflow.

🎯 AI Auto-Zoom

✔️ Detects clicks automatically

✔️ Applies smooth cinematic zoom effects in real time

✔️ No manual keyframing required

⚡ 12x Faster Export Engine

✔️ Export 4K videos in seconds

✔️ No re-exports

✔️ No bitrate tweaking

🎥 Pro Webcam Overlay

✔️ Circular facecam overlay

✔️ 4 placement options (Top Left, Top Right, Bottom Left, Bottom Right)

✔️ Hybrid-Sync technology keeps audio and video perfectly aligned

🎨 37+ Studio Backgrounds

✔️ Professional gradients and studio-style scenes

✔️ Instantly improve the look of recordings

✂️ Built-in Editing

✔️ Trim recordings

✔️ Crop videos

✔️ Export without opening another editor

🔍 Custom Zoom Controls

✔️ Zoom intensity from 1x to 4x

✔️ Disable Auto-Zoom anytime for full manual control

📱 1-Click Vertical Export

✔️ Export directly to 9:16 format

✔️ Perfect for Shorts, Reels, and TikTok

🔒 Privacy First

✔️ No account required

✔️ No watermark

✔️ No cloud uploads

✔️ Everything runs locally on your machine

💻 Cross-Platform

✔️ Works on Windows and Linux

✔️ Not limited to Mac users

💰 Pricing

🆓 Free tier

✔️ Fully functional

✔️ No watermark

✔️ No signup required

🚀 Lifetime Access

✔️ $9.99 today

✔️ Started at $4.99

✔️ Will increase to $29 in the next version

✔️ Only 5 spots remaining at the current price

🔗 Links

👉 Install Free: Chrome Web Store

🌐 Landing Page & Demo: https://snapzoom.pro

I'd love honest feedback:

❓ What's missing?

❓ What's confusing?

❓ What feature would make you use SnapZoom every day?

I'll be here all day answering questions and collecting feedback. Thanks for checking it out! ❤️

Hey PH 👋

We built Outbound Rewriter because we were sick of "AI email writers" that do this:

1. Ask you to paste a bio

2. Spit out a generic paragraph full of "revolutionary" and "game-changing"

3. Ghost you when it gets zero replies

So we did the opposite.

Drop in a LinkedIn URL. The tool scrapes the profile, builds an Evidence Wall of talking points (metrics, pain signals, growth indicators), picks from 5 psychological frames, and writes a message that actually sounds like a human sent it.

Not a template. Not a prompt. A researched, angled, scored piece of outbound.

What you get:

🔍 Automatic prospect research — no more copy-pasting bios into ChatGPT
🧠 5 psychological frames — Pain Echo, Trigger, KPI Mirror, Social Proof, Curiosity Gap — scored by predicted reply rate
🚫 Anti-cringe linter — real-time buzzword and spam-trigger detection
📝 Reply-state branching — first touch, follow-up, or "not interested" recovery
📊 Reading-level + compliance scoring — Flesch-Kincaid + GDPR/CAN-SPAM built in
🎭 Voice cloning v2 — upload your best writing samples, the AI learns your tone
🔄 "Regenerate with coaching" — annotated rewrites that teach you why it works
✏️ Inline editing + tone shifts — tweak any line, shift formal → punchy in one click
👀 Deferred lead gate — see a full before/after preview before you unlock anything. No bait-and-switch.

The bigger picture

Outbound Rewriter is free. And it's one of the 4 free AI tools we're releasing:

We built these so any founder or SDR can sharpen their GTM motion without paying enterprise prices. Use them individually, or connect them into a scalable AI workflow with AI Engage™ — our full revenue operating system.

Try it free

👉 knowledgenet.ai/tools/outbound-rewriter

I'd love to hear what you think, especially if you've tried other AI email tools and felt let down. I'll be in the comments all day.

And when you're ready to move beyond one-off personalization and automate engagement at scale, it's time to explore AI Engage from KnowledgeNet.ai. AI Engage handles the entire process hands-free, helping you personalize and manage outreach across large audiences without sacrificing authenticity.

KnowledgeNet.ai is opening some of its AI tools so you can use them for free and learn. If you like to operate at scale, connecting multiple AI toolkits to create a scalable AI workflow, book a meeting with our experts.

Okay. This is actually unique. Do you have career coaching on your website.

The dream use case: open laptop, pick a month, find where the sun is - and just go. Post-COVID remote work made this kind of spontaneous travel actually possible for a lot of people, and tools like this make it feel less like planning and more like choosing

Built this after a single bad prompt quietly doubled my Anthropic bill and I had no idea which deploy caused it. Helicone and Langfuse are great but they show you the symptom, not the cause.
LLMTrace is a Go proxy that tags every LLM call with the commit SHA that triggered it. Self-hosted, no vendor lock-in. Happy to answer any questions about how the attribution works.

Congratulations on your launch! Being able to trace usage back to the exact deployment while keeping everything self-hosted makes the debugging process much more practical and trustworthy.

swain uses a fixed severity ladder: launch-risk → high → medium → ok. anything that could get you breached on day one (billing bypasses, auth holes, hardcoded secrets) sits at the top.

if two findings share the same severity, it goes by category, billing and auth before XSS before tenant isolation, because that's roughly the order a motivated attacker works in.

How does swain decide what to prioritize when it finds multiple potential problems?

The rich-tag idea is strong because it avoids the usual journal-vs-database tradeoff. A lot of life-tracking apps make you choose between freeform notes that are hard to analyze and structured fields that feel too heavy to keep using.

The feature I’d be most careful with is the AI summary layer. For personal journals, summaries need to preserve “why this mattered,” not just compress events into a clean recap. I’d want controls like: summarize only selected tags, show which entries influenced the summary, and let me mark a summary as wrong so the system learns what not to flatten.

That would make the insights feel personal instead of generic.

Hi PH community!

I’m Nicola, the creator of TagJournal.

I’ve always struggled with finding the 'perfect' app to track my life. Most tools are either too specific (just for fitness or just for finance) or too complex. I wanted something that felt like a journal but acted like a database: fast, flexible, and visual.

TagJournal is a web and iOS application designed to help you effortlessly track what matters most.

For years I used a web app I had built for myself in my spare time (in .NET!) where I could add tags to my days to track basically anything and then get simple statistics out of it. At the same time, I was journaling using another app.

I always had many ideas for improving that application, but never enough time to work on it. AI helped accelerate that process, and in just a few months I finally built the app I had always wanted.

The idea is simple: using rich tags with flexible metadata.

The smallest possible entry you can add to a day might only contain date, time, and the tag name. But for each tag, the user can decide whether to enable these metadata fields:

• Icon

• Long text (editable later with a distraction-free editor), useful for journaling but also for many other types of content

• Numeric value + unit of measure (which can also work as a 1–5 rating)

• Currency

• Duration

• Photos

• Location / Place (using Google autocomplete)

• Person (I needed this to track information related to other family members too)

Tags are divided into color-coded categories.

This structure allows different kinds of objects/features:

Tag Watch: select a tag, a time period (current/previous week, current/previous month, current/previous year), and the field to monitor (duration, value, or currency, either total or average). The user can then keep that metric always visible on the “Today” page.

Goals: same idea—select tag, period, and field, then set a target. In “Today” you always see the completion percentage.

Charts / Trends: similar to Watch, but data is displayed as charts. These can also be added to the “Today” dashboard.

This way I track everything I need to monitor or review later:

• Symptoms, medications, medical visits, physical activity, food and drink consumption, mood

• Work activities, meetings, people I meet, focus sessions, releases

• Information related to my children such as grades, absences, sports activities

• Places, food, books, movies, TV shows, apps, things I learned

• Portfolio performance, income, expenses, and specific spending categories

This is the core concept, but there are many other features too: map view for location-enabled tags, photo gallery, calendar, collections, etc.

Another thing I always wanted, and can finally use, is the Zapier integration to create entries from other platforms (for example, I track GitHub commits or calendar events). It’s also possible to create custom integrations by generating API keys directly inside the app.

Why I’m launching today: TagJournal is currently in a phase where validation is key. I’ve built the core features, rich tags, AI-powered summaries, and advanced statistics, but now I need your eyes on it. I want to know:

• Does the 'tag everything' approach resonate with your workflow?

• Which features (AI assistant, Goals, or Zapier integrations) are most valuable to you?

• What’s missing to make this your daily driver?

I’ll be here all day to answer your questions and take notes on your feedback. Thank you for helping me shape the future of TagJournal!

Happy tagging!