PH热榜 | 2026-03-13

he SKILL.md format has quietly become a lingua franca for how builders encode their best thinking into AI agents.

Claude Code popularized it.

Codex adopted it.

And now Perplexity Computer supports it which means the workflows you've spent months refining don't have to live in one tool anymore.

That's the actual news here. Not just "Perplexity added a feature." It's that your institutional knowledge the how-to-handle-this, how-to-structure-that logic baked into your skill files is now portable.

The gap this fills:

Most power users of Claude Code or Codex have accumulated a small library of SKILL.md files.

Presentation builders.

Research frameworks.

Weekly briefing templates.

These aren't throwaway prompts they're curated playbooks that encode real judgment.

Until now, those lived in one tool's ecosystem.

What Skills actually does in Perplexity Computer:

• Upload any existing SKILL.md file directly -- it works as-is if it has the right YAML frontmatter

• Describe a workflow in plain language and Computer builds the skill for you (no technical knowledge needed)

• Skills activate automatically based on query context -- you don't have to invoke them manually

• Computer can combine multiple skills mid-task (Research + Content for a blog, for example)

• Browse and manage your library from a dedicated Skills tab

The execution layer underneath is what makes this different from just "importing a prompt."

Your skill now runs on 19 specialized models: Opus 4.6 for reasoning, Gemini for deep research, specialized models for images and video.

A SKILL.md that told one model what to do now tells the best model for each sub-task what to do.

Who this is for right now:

Builders and indie founders who've invested in building agent workflows on Claude Code or Codex.

If you've built more than five SKILL.md files, this is probably worth your attention today.

If you're newer to the format, Computer can generate skills from a plain-language description, so there's an on-ramp.

One honest caveat:

Computer is currently Max-subscriber-only at $200/month.

That's not nothing.

Pro and Enterprise rollout is coming, but if budget is a constraint, that's the real friction point to weigh.

The bigger question I keep turning over: as SKILL.md portability becomes table stakes, does the agent tool that wins end up being the one with the best execution layer or the one with the best skill library ecosystem? Perplexity is clearly betting on the former. Curious whether anyone here has already migrated workflows over and what the quality delta looks like in practice.

Amazing am gonna try this out

The SKILL.md portability angle is the real story here. I maintain a library of 100+ skills for Claude Code and the biggest friction has always been vendor lock-in. Being able to reuse those workflows across tools without rewriting is a genuine time saver. Curious how it handles complex multi-step skills with tool-specific syntax.

@rohanrecommends @aravindsrinivas @Perplexity Computer Skills As a heavy user of AI workflows, this "Computer Skills" module is really quite different.

Upgrade "prompt" to "reusable workflow template

Previously, long prompts were piled up in various chat interfaces, which were difficult to reuse and could not guarantee stable output. Perplexity's Skills transform these frequently used patterns into "naming and manageable skills", and they can be triggered automatically based on descriptions, providing a more experience similar to "requesting a fixed style AI employee" rather than randomly generating a prompt.

The two creation methods cover users ranging from beginners to advanced users.

If you can't write the SKILL.md file, you can directly generate it in the Computer using the "Create with Perplexity" dialog box. This is very suitable for product, operation, and analysis students.

For teams with their own knowledge base and standards, you can manage it using a .zip + SKILL.md format. Combined with the YAML frontmatter's name and description, it can achieve precise triggering and is already very close to the feeling of "lightweight workflow orchestration".

The automatic activation driven by the description + combined use is the highlight.

The series of relay tasks mentioned in the document, such as Research, Research Report, and Slides, are very crucial:

Users only provide requirements, and Computer selects and arranges the skills based on the descriptions.

In the same task, one can first conduct research, then write a report, and finally generate a presentation. This minimizes the need for manual copy-pasting.

This is closer to the real knowledge work scenario than a "super universal chat box".

It is very friendly to team collaboration and standardized output.

The "weekly report template", "competitive product research paradigm", and "data analysis report structure" are all made into Skills. New team members can directly reuse the SKILL.md accumulated by experienced members, which not only saves training costs but also ensures a more unified output style.

The following are the several points that I will pay attention to.

Permissions and Sharing: How to manage within a team or enterprise who can create, who can modify, and who can share skills

Version Control: SKILL.md, as a text, is actually very suitable for working with Git for version management

Observability: When complex tasks involve multiple Skills and the result is unsatisfactory, can one see the "call chain" to facilitate optimization

Perplexity has quickly become one of my favorite tools for researching things online: getting answers with sources instead of digging through dozens of links is such a better workflow. Congrats on the launch! What feature are you most excited to build next for power users?

Computer use is the big next step. Going from "answer my question" to "do this task on my computer" changes everything. I've been waiting for this since Anthropic demoed computer use last year. The repeatable instructions part is what I care about most. I have about 15 things I do every morning that could probably be automated if the AI could just watch me do it once. How reliable is it with apps that change their UI frequently? That's usually where automation breaks.

The SKILL.md portability across Claude Code, Codex, and now Perplexity is a smart move — making workflows tool-agnostic rather than locking them into one ecosystem solves a real pain point for power users. Curious whether there's a versioning or diff system planned for skills, since iterating on a workflow across 19 models likely produces very different outputs than the original single-model setup.

This looks really interesting.

Are the repeatable instructions more like automation workflows or closer to scripting for the computer?

Curious how flexible it is for developers.

Google Maps is reimagining navigation with Gemini.

The new Ask Maps feature lets you ask real-world questions conversationally and get personalized recommendations powered by data from over 300M places and insights from 500M+ contributors.

Instead of searching multiple tabs, you can simply ask things like where to charge your phone, find a tennis court at night, or plan stops on a road trip. Maps then shows options on a customized map and lets you take action by booking, saving, or navigating.

Google is also introducing Immersive Navigation, the biggest navigation update in over a decade. It adds vivid 3D route views, highlights lanes, traffic lights, and crosswalks, and gives more natural voice guidance to help drivers prepare for turns, merges, and exits.

The result is a more intuitive way to explore places and navigate routes with real-world context and smarter guidance.

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

This feels like the natural evolution of maps.

Is the AI answering queries using real-time map data or more like a conversational layer on top of existing search?

Super curious how navigation + AI interaction will blend here.

Its not clear to me how you can opt in to this functionality and use it?

The shift from keyword-based place search to conversational queries like "where can I charge my phone nearby" changes how people interact with maps entirely — it's less about finding a pin and more about solving a real-world problem in context. How does Ask Maps handle ambiguous or subjective questions like "best quiet café for working" — is it pulling from reviews sentiment, visit frequency, or something else?

Has seen this comparison (Google vs Apple 2021). Dunno, Apple still feels a bit user friendlier: https://x.com/techdroider/status/2032146599783383196?s=20

Simple yet effective. Tools that make other tools better are always a plus! Does it use real time map data to answer questions, etc?

Interesting expansion! Immersive navigation seems fantastic. I assume the corpus here for a lot of the conversational elements is going to rely heavily on GMB information, which, for local businesses and discovery, will emphasize even more the importance of a fully fleshed-out GMB profile.

As someone who's driven 2,000+ miles with Maps navigating, I'm curious how the LLM handles the ambiguity when someone asks "find me a coffee shop" while you're actively navigating—does it prioritize upcoming exits or reroute you somewhere better? The demo video's voice recognition seemed suspiciously good at parsing "somewhere quiet with good espresso near my route" without follow-up questions.

"Planning is not review. Review is not shipping. Founder taste is not engineering rigor. If you blur all of that together, you usually get a mediocre blend of all four."

I resonate with this so hard. Agents (like humans) can't perform as well when they wear too many hats. The future is coordinating teams of specialized agents like this with just the right amount of human in the loop.

That means you can have one session running /qa on staging, another doing /review on a PR, a third implementing a feature, and seven more working on other branches. All at the same time.

It's a freeing feeling when your workflows reach this level of parallelization. The productivity gains are insane.

Can't wait to try these out in Birdhouse. Thanks for sharing!

I've been using Claude Code with MCP for about a year now and keep iterating on my setup. Curious what's in Garry's stack — is it mostly about the MCP server configuration or also specific prompt engineering patterns? The biggest unlock for me was getting file system + browser MCP tools working together. Does GStack cover the agentic workflow side or is it primarily IDE-focused?

The slash command approach resonates with how we've been running Claude Code autonomously. We built 100+ SKILL.md files over 108 hours of unattended operation and the key insight was exactly this: specialized roles outperform a single generic agent. The biggest win was separating "plan review" from "code review" — when Claude tries to do both in one pass, it either rubber-stamps or gets lost in details. Question for the team: does GStack handle context window pressure when running multiple skills in sequence? That's been our main scaling challenge.

Why is this on Product Hunt?
This is a bunch of prompts. You have got to be kidding.

Splitting Claude Code into role-specific slash commands like plan review, code review, and shipping is a practical pattern — it enforces structure that most developers skip when using a single generic agent. How opinionated are the workflow skills under the hood — are they rigid step-by-step procedures, or do they adapt based on the codebase context and project size?

Just got here to support the legend Garry who’s now back to coding like he’s 21. Keep up the great momentum G!

This is amazing, Garry!

I am curious how you use it. do you usually run things sequentially, or is it parallel most of the time? And in the parallel case, how do you keep the flow in sync when each sub-agent is continuously making changes- do you use git worktree?

can't wait to dive into this this weekend

Fastest easiest path to level up with CC

Congrats on the launch! GStack is exactly what the Claude Code ecosystem needed — role-based specialization instead of one generic mode.

One thing worth adding to this setup: a seatbelt layer.

We ran GStack-style autonomous sessions and hit Case #001: a Claude Code agent looped for 70 minutes, repeatedly injecting a staging URL into a production config file. Every log showed exit code 0. All green. The deviation was invisible because nothing had recorded what the agent intended to do before it acted.

Built K9 Audit to fix this — a deterministic, non-LLM causal auditing layer that drops directly into .claude/settings.json (zero code changes, fully compatible with GStack). It records a cryptographically hashed 5-tuple of declared intent vs actual outcome. When something goes wrong, k9log trace --last gives root cause in under a second.

GStack for velocity. K9 Audit as the seatbelt.

Repo: https://github.com/liuhaotian2024-prog/K9Audit

Cool to see you launching this! Very useful too - I've already pulled in a few of these for my setup.

love the name xD

I've been using Claude Code for months and my setup is held together with scattered markdown files and random CLAUDE.md instructions. The idea of a curated, tested configuration from someone who actually pushes it hard is appealing. My main question: how opinionated is this? Half the value of Claude Code is customizing it to your specific codebase. If this overwrites my existing CLAUDE.md and memory files, that's a dealbreaker. Does it layer on top of existing config or replace it?

the slash command approach is really smart. I've been using Claude Code for a while and the biggest friction is always starting from scratch with context every time. having pre-built workflows for common tasks like code review and shipping saves so much time. curious if you're planning to add custom skill creation so teams can build their own workflows too?

Love seeing developer environment setups shared like this.

Is this mainly optimized for Claude Code workflows or does it also work well with Cursor / other AI coding tools?

cool! easy setup! thanks gary

Hey Product Hunt, I'm Joel, co-founder of Pinnacle!

The Problem

Our world today is overwhelming. Always-on culture and increasing demands have created an epidemic of stress, fatigue, distraction and burnout. Yet the tools available aren’t built to help people perform at their best under daily pressure. 

• Therapy is incredibly valuable for mental health support, but it's expensive and geared more towards processing past experiences rather than future-focussed personal growth.

• Coaching is more relevant to performance enhancement but still expensive, decentralised and lacking quantified measures to track progress.

• Meditation apps teach profound mindfulness practice, but these experiences are impersonal and static, lacking adaption over time or a deep understanding of the user.

• Smart Wearables measure your body but don’t measure your mind. Insights are limited to physiological indicators, lacking a more complete psychological picture.

The Pinnacle Solution

Pinnacle turns your iPhone into a tool for understanding and training your mind.

• Using the sensors already built into your phone, our AI reads real-time biometrics, interprets your emotional state, and pulls everything together into a clear picture of how you're actually performing.

• Instead of raw data, you get simple, actionable insights you can use throughout your day. With Pinnacle, you can track and improve your focus, HRV, emotions, sleep, and overall performance.

• Think of Pinnacle as a mirror for your inner world. It surfaces patterns you wouldn’t normally notice and helps you make small, meaningful changes that add up over time, so you can perform better day after day.

• For the first time, if you have an iPhone, you can truly quantify and upgrade your mind.

Who is this for?

If you're a knowledge worker experiencing high pressure and feeling stressed, fatigued or distracted, Pinnacle helps you build resilience and thrive under pressure. Upgrade your brain with Pinnacle.

Get started today

From today, the Product Hunt community gets access to Pinnacle Core for free. Try Pinnacle at pinnacle.co and see what you can achieve.

In the future, Pinnacle will start charging a monthly subscription. If you love Pinnacle, please message us directly with PHLOVE. The first 50 folks to do that before Saturday March 14th will get their first year of Pinnacle Core for free.

Huge fan of Pinnacle. Genuinely one of the best UXs I’ve seen

Nice launch! Tools that focus on improving real workplace interactions are always interesting to see. I’m curious how the AI balances providing guidance while still keeping conversations natural and unobtrusive.

Using the iPhone's built-in camera for HRV measurement without requiring a wearable removes a huge adoption barrier — most brain performance tools fail because they add hardware friction before delivering any value. With the attention tracking feature coming soon, how are you thinking about privacy given that it likely requires continuous camera access during work sessions?

Love the concept. I’ve tried prompting chatgpt to be my coach but it hasn’t led to great results. What’s your approach that’s differentiated from existing LLMs?

have been an early user! I'm impressed with how far its come. The interface and interaction is sci-fi and love the HRV and emotional detection features. What's next on the roadmap?

So it's like Whoop for the brain? Love the idea. This will go big. Turn your phone from brain rot to brain healer. Good luck team!

Interior designer and working mum here—my life is basically a constant battle against physical and mental clutter. I’m naturally skeptical of 'performance' tools because they usually just add more noise to an already loud day.

Pinnacle feels different. From a design perspective, the spatial hierarchy and use of negative space is calming, like walking into a clean studio after a chaotic morning with the kids. Was 'mental load reduction' a core part of the brief, or did the minimalist UI just happen to solve that for us busy parents? Love using the app so far!

I’ve been using EEG-based feedback apps for years, and the friction of carrying a headset kills the habit. If Pinnacle can really pull cognitive-state detection from nothing but the IMU and camera I already stick in my face every morning, that’s a 100x convenience win—would love to hear how you validated signal fidelity against a clinical-grade device.

@joel_jackson1 Congrats on the launch! The angle of using sensors already on your iPhone (no wearable required) is a winner for me. Curious: how are you handling the calibration challenge? iPhone camera-based HRV readings can vary a lot based on placement, lighting, and stillness. What does your accuracy benchmark look like ?

Turning a phone into a brain performance coach is a cool idea.

Are the exercises based on neuroscience research or more behavioral training?

Would love to know how the progress tracking works.

Been using Pinnacle for a few months and it's one of the few (if not only) products which has a demonstrable impact on focus 😀

Big fan and hope many more people improve their lives through this launch!

Congrats on the launch!

This looks great! Does it have a fairly easy ramp up where you can try a small thing for a while and feel its impact before trying more?

Why do only the iPhone users get the good stuff?! 😭

Any thoughts on how Pinnacle can help deal with phone addiction?

Sounds simple, but my eyes by default feel "tired" how does Pinnacle adapt to understand that's my normal eye expression? Congrats on the launch!

Really interesting concept: turning your phone into a brain performance coach using real-time signals sounds super futuristic. Congrats on the launch! Which sensor or signal has been the most useful for understanding someone’s focus or mental state so far?

Congrats on the launch, Joel! Really impressive to see how you're leveraging on-device sensors for these kinds of emotional insights. The focus on performance and resilience for knowledge workers hits home.

Quick question: what sensors on the device are you using?

Would love to see how this evolves. Upvoted!

Looks really cool.
I in France and it doesn't seem to be available here. Will it be anytime soon ?

Hey! Can Paul handle hiring for multiple roles at once, or just one at a time?

Framing recruiting as a recommendation engine problem rather than a search problem is a meaningful distinction — the continuous calibration loop from interview feedback is exactly how TikTok's content matching works, and it makes sense applied to hiring. How does Paul handle the cold start problem for a new client with zero historical feedback — does it bootstrap from the voice brief alone, or is there a broader signal it pulls from?

Really interesting idea: replacing the traditional recruiting pipeline with an AI agent that handles sourcing, outreach, and screening could save startups a huge amount of time. Congrats on the launch! How does Paul decide which candidates are the best match after the initial sourcing and screening steps?

Sometimes the best candidates are the ones that aren't actively searching right now but could be potentially "poached". Does Perfectly consider these types of candidates as part of its "Outreach" process? Or are the candidates only the ones that are actively searching?

The “recruiting tax” framing hits hard — every founder I know has lost weeks to sourcing, screening, and chasing candidates

Question: at what company size does Paul perform best right now? Curious if the calibration model needs a certain volume of interview feedback before the matches get really sharp, or if even a 3-person team can get value from day one.

Congrats on the launch!

That fork like the microphone holder. I can't. 🤣

"Voice-to-Stack" sounds like a massive time-saver. Does Paul actually draft the JD based on that 5-minute brief, or just use it for internal search parameters?

It might be helpful to add reminders for when to reconnect with someone after the first msg.

Congratulations on your launch. Beyond the initial outreach, does Parker help track the status of these conversations, or is it primarily focused on getting the "foot in the door" initially?

If there isn't an active role available but it's at a company that I'd be interested in joining, are there ways that Parker can either (a) monitor for when a relevant posting appears or (b) help me craft outreach that could get me in the door even though there isn't an active job posting?

Meeting people where they already are — iMessage and WhatsApp instead of yet another app — is the right distribution call, since the 70-80% referral stat means the real leverage is in warm intros, not cold applications. How does Parker differentiate between a genuine networking opportunity and a connection that would feel spammy to the recipient — is there a signal quality threshold before it suggests reaching out?

Congrats on this @victor_luo The iMessage + WhatsApp angle is underrated. Every other AI career tool puts you in yet another dashboard you have to remember to open. When Parker drafts personalized outreach that ‘sounds like you’ — how does it actually learn your voice? Is it from the iMessage/WhatsApp chat history alone, or does it analyze LinkedIn activity, past writing samples, etc.?

How does Parker handle data privacy? Mapping out who to talk to sounds powerful, but I want to ensure my professional network data stays completely secure.

As someone who just spent 3 hours debugging a deployment because our junior dev didn't understand the difference between staging and production variables, I can see why you'd want to abstract away the operator role entirely. The real question is: how do you handle the edge cases where someone needs to SSH into a box at 3am because the automated rollback failed, or when the "simple" deployment requires understanding the intricate dependencies between your Redis cluster and PostgreSQL replicas?

The no-edit Friday report sent to someone you don't want to disappoint is a clever forcing function — most accountability tools fail because founders can reframe their own narrative before sharing it. Does the AI advisor push back in real-time when you set weekly goals that are clearly too safe or disconnected from your North Star, or does it only surface that gap in the weekly report?

I've been watching this product since its early concept days and I'm super impressed with the newest version. It has always provided accountability for founders, but this latest version makes it so much easier to get going – and keep going. Great work!

How does the system get context with regards to where I am at in my startup journey? Are they solely based upon the "north star" I set or are there ways to get more personalized feedback once it understands the industry and product that I'm trying to sell?

This is launch of the week for me. Love this product and plan for our team to use it.

Love seeing all the progress, Bubs! This is such a great idea and needed product. Good luck with the launch!

Really interesting concept: an AI chief of staff that tracks your real progress and sends honest weekly updates could be huge for founders trying to stay accountable. Congrats on the launch! How does Pre decide what data signals actually matter when measuring whether someone is making real progress toward their North Star? 🚀

Last week I got hit by a client with "sorry we took all the docs work your team did over the last 3 months which was great, fed it to Claude Code and we're good going forward". $5k+ MRR up in smoke.

I think that's when I might have finally gotten past the denial stage, that AI is coming for my company, Hackmamba, a technical content agency, even though we're focused on authenticity and technical creativity.

As an engineer and technical writer (now double-screwed I guess) I'm a big purporter that AI is like electricity, making things better, but the last 2 weeks have been, shocking (pun intended). Maybe I'd just been slow, doing too much talking and less doing.

So what did I do after J hit me with the contract cancellation line, I started looking for ways to do more with AI without crossing the blurry line that is generating slop. As a former PM, the first culprits of my evaluation were anything we spent more than 10 hours per month doing.

Technical reviews came up first. We work in teams shipping fast and need to get docs ready for developers and agents. Documentation is the ground truth before MCPs etc take over. So we spend a good amount of time reviewing docs PRs sent in by technical writers for accuracy, tone, shit code, typos, consistency with the overall style, persona match, clarity for sales and marketing usage etc.

So I did the next logical thing a software engineer (bless that job title) would do; I made a system prompt with everything we know and documented internally, plus everything I know about docs, individual frameworks, patterns etc. Then I built Fowel.ai (should sound like vowel, not foul) with it to handle deep GitHub PR reviews on documentation that was both written by a human or AI generated.

Frankly, I don't care at this point. If the end goal is to ship great docs for humans and agents, why care deeply about who wrote it. AI agents don't care, and I've worked with writers worse than GPT 5.4. We likely won't need documentation in the future too when we fix agent<>agent comms 🫠

Maybe I'm cooked for making such mental shift towards building the guardrails and quality enforcements. Time will tell.

We've seen a huge reduction in time to get PRs into production by about 80%, which I love. Do try Fowel if you're looking at the speed of getting great docs content out, and I appreciate any feedback shared. It's free to use.
Thanks in advance and let me know if this is shit too.

I don't mind brutal feedback.

William.

This is honestly such a smart pivot. Doc review is one of those things that eats up so much time but nobody really talks about it. We've had PRs sit for days just because the docs changes needed back and forth on tone and accuracy.

Curious about one thing though, does Fowel handle docs that reference external APIs or third party libraries? Like if a code sample imports something from a package that just shipped a breaking change, would it flag that?

The agency-to-SaaS pivot triggered by a client replacing your work with Claude Code is painfully honest and exactly the kind of signal that validates the product direction — if AI can write the docs, the value shifts to reviewing and enforcing quality standards. Does Fowel learn from a repo's existing documentation style over time, or does each PR review start from a generic baseline regardless of how much context already exists?

This resonates hard. I maintain 440+ open-source developer tools and the docs review bottleneck is real — outdated code samples slip through PRs constantly. The fact that Fowel catches structural gaps (not just typos) is the key differentiator. Most linters stop at formatting. Curious: does it handle multi-language docs, like when a project has both English and Japanese README files?

Since I rely so heavily on Claude to help me with development work these days, I find myself questioning if the documentation that I write up is detailed enough or if I missed a key detail.

Is Fowel able to build upon context across different repositories that share dependencies?

An 80% reduction in doc review time is ambitious, but I'd accept it for certain use cases. Most of the time reviewing docs is spent on consistency checks and finding outdated references, and that's tedious work a human shouldn't be doing. I maintain docs for a SaaS with 5 language localizations and the drift between versions is constant. Does this handle multilingual documentation or just English?

The 80% review time reduction tracks with what I've seen when you encode institutional knowledge into the prompt rather than relying on generic LLM behavior. How are you handling style guide drift over time? Like when a client updates their tone or deprecates certain terminology, is that a manual prompt update or do you have a way to version that context?

The personal assistant angle — asking "what are my travel plans" and getting a consolidated view from booking emails — is a much stronger value prop than just inbox sorting, because it turns email from a chore into a queryable knowledge base. How does Replyless handle multi-account setups where someone has work and personal email with very different categorization needs?

This looks really interesting, but I'm curious about the quality benchmark — in what specific scenarios does ReplylessAI outperform Superhuman, and where would you honestly say it's not there yet?

One of my biggest issues is that I subscribe to too many newsletters that I mean to read but never have time for. Are there ways for Replyless to help generate summaries and pull out key insights from an aggregation of newsletters from the week?

Congrats @srivatsa_mudumby This is real time saver. Quick question: how does Replyless handle false positives? If the AI miscategorizes something important as noise, is there a feedback loop to help it learn your preferences over time?

Embedding a full AI agent inside Telegram instead of building yet another standalone app is a smart distribution play — Telegram's bot ecosystem already trained users to expect automated workflows in chat. What's the latency like for multi-step tasks compared to using Manus directly, and are there any capability trade-offs from running through the Telegram interface versus the native platform?

Hi everyone!

I built this from my own apartment hunt needs, as I was surprised that there are no similar tools online at all! With this tool, I was finally able to check exactly which floors and units of the apartment building have satisfactory view, privacy and sunlight, without having to guess.

Now, for your apartment hunt, you can:

• 👀 See the outside view from a specific floor or unit.

• 🏙 Check obstructions by other buildings.

• ☀ View year-round daily sun exposure.

• 🔗 Send the exact 3D view to roommates or brokers via URL.

Window View is powered by Google Maps Photorealistic 3D Tiles (through Cesium ion) and CesiumJS. Feel free to contribute or open issues for any features you want to see! https://github.com/wengh/window-...

So cool. How do you plan to monetize it? Congrats on the launch!

The shareable URL for roommates and brokers is really helpful. Are you planning to integrate directly with listing platforms like Zillow or Realtor.ca ? Also a “compare units” mode where you can pull up two apartments side by side would be cool. Congrats on this!

This is one of those “why didn’t this exist already” products. Renting an apartment and being burned by a listing photo that conveniently hid the parking garage wall outside the window? Not ideal. Congrats on your launch!

This is genuinely impressive. I just spent way too much time playing with it—it feels more like an immersive travel experience than a utility tool. Using Google Earth 3D tiles for something as practical as checking apartment views is a brilliant execution. Does it currently support sunlight simulation for specific seasons, or is it based on the current date?

Really interesting idea. I like the concept of checking the view before renting.

Curious: how do you handle cases where Google Earth data is outdated or buildings have changed? Does the tool account for that in any way?

So proud to see this go live! 🎉 The team has been shipping non-stop since February — Database Blocks, Inline Comments, Multi-Column Layouts, AI Thinking Mode — every feature came from real user feedback. Excited to see what the PH community thinks. Happy to answer any questions! 🙌

Tried a bunch of "AI writing tools" this year and they all feel the same -chat window on the left, doc on the right, hope for the best. The skill system here is what's actually different. Great job on the UI!

Hey @wangzhang_wu Congrats on this! With thinking Mode now handling complex reasoning, how are you approaching context windows when the AI is referencing multiple database blocks and documents simultaneously?

“8 updates in 6 weeks” is the kind of shipping cadence that tells you everything about a team’s conviction. Congrats on this @wangzhang_wu Are you planning a Notion import feature so users can migrate their existing workspace?

Looks cool, but I wonder how companies feel about candidates using an AI overlay during interviews, might be a tricky line.

How is it different from old Cluely? Congrats on the launch!

As someone who's had to awkwardly pause screenshares to search docs mid-call, the invisible overlay concept is brilliant. Being able to silently pull up relevant client data or talking points without breaking flow would save me from those "let me just..." moments that kill momentum.

TThhee iinnvviissiibbllee--ttoo--ssccrreeeennsshhaarree aapppprrooaacchh uussiinngg aO SW-ilnedvoewls oovveerrllaayy rtehnadte rbiynpga sisse st escchrneiecna lclayp tculreev eArP I— sm oisst tAeIc hmneiectailnlgy acslseivsetra n— tmso satr eA Ie imteheetri nvgi saisbslies ttaon tesv earryeo nvei soirb lree qtuoi reev etraybo-nsew iwthcihcihn gl itmhiatts bhroenaeksst yuosuarg ef.l oCwu.r iCouursi oaubso uatb otuhte tehteh iectahli cbso uanndgalrey: hheorwe :d od oy oyuo us esee ec otmhpiasn ipersi maanrdi liyn taesr vai epwreerpsa rraetsipoonn dtionogl afso rt hheisgeh -tsotoalkse sb eccoonmvee rmsoartei ocnosm,m oonr, aarned uisse rtsh earles oa rmuondnei ntgh aitt mtaok egse tG hroesatlD-etsikm ev iasnisbwleer sf odru rtirnagn slpiavree ntte cuhsnei ccaals eisn tleirkvei eswasl?es prep?

@mit_parikh Congratulations on this! I really resonate with the no tab switching. Curious how GhostDesk handles heavy accents or fast speakers? Would love to know what your real-world WER looks like in noisy call environments.

Congrats @mit_parikh I appreciate the ease and accessibility of this! Does GhostDesk support custom knowledge bases? For example, could I feed it my product docs, or research notes so it surfaces my specific answers rather than generic GPT responses?

"Hey Product Hunt! 👋 I’m Bryan, the maker of ClawMote.

I built this because I was tired of being 'desk-tethered.' I’d be leaning back on the couch, deep in a Wispr Flow dictation or an OpenClaw session, and I’d have to physically reach forward just to hit that FN key or keyboard shortcut or move my mouse just to get the pointer in the chat window. It was the only thing stopping a truly free AI workflow.

ClawMote is the bridge. It’s a lightweight, set-and-forget menu bar utility that puts total Mac authority in the palm of your hand.

Why you’ll love it:

One-Hand Total Control: Map Wispr Flow, OpenClaw, and keyboard shortcuts to a handheld device.

Lightweight: It lives in your menu bar and stays out of the way.

No Subscriptions: Try it free for 7 days. If it changes your workflow, it’s a one-time $20 payment to own it forever.

I’m here all day to answer questions. I’d love to hear—what’s the one 'desk-bound' shortcut you can’t wait to move to your handheld setup? 🦀🦾

I know it's simple, but I have Command+S mapped to a button for quick and easy OpenClaw session saves as backups!

Ditch the desk. Keep the power."

🚀 Community Spotlight: Shoutout to@jerrybyday for the great question on custom terminal commands! To clarify for everyone: ClawMote is system-wide and can be used with any app on Mac.

Congrats on this @bryan_lee21 Any plans to expand the mappable apps beyond Wispr Flow and OpenClaw? As someone building with AI tools daily, I’d love to see support for custom terminal commands down the line.

The privacy question Liora asked is the one I'd want answered before keeping this open all day. Local processing is a dealbreaker for a lot of people, and if you're doing it on-device, that should be front and center on your landing page, not buried in a FAQ.

it will be great if the app could track your posture via webcam and send you notification when you go on the orange/red zone (bad posture). Good luck!

As someone who's been working from the same desk for 3 years and just paid $800 for physical therapy, I'm curious about your privacy model - is all the posture detection happening locally or does my slouching data make a round trip to your servers? The real-time feedback loop sounds perfect for preventing the slow creep of bad habits that only hurt after months of reinforcement.

I appreciate the “no wearables, no hardware, nothing to install” Congrats on achieving this! Any plans for a focus mode integration, where SitSense only monitors during active work sessions and pauses during breaks? This make it more like a coach rather than surveillance.

Hey PH! I'm Shalin, building Hyper with my co-founder Kanyes.

At our last company (a robotics startup), the most important decisions never happened in scheduled meetings. They happened spontaneously (in hallways, over lunch, on walks). None of it got captured. And every recorder we tried was built for Zoom calls with agendas, not for real life.

What we realized was that even the tools that DO capture your conversations give you a document, a summary, or notes as the final artifact. And then what? You file it somewhere, never open it again, and three weeks later you're in another meeting going "wait, what did we decide about the deadline?"

We think the whole paradigm is wrong. You don't want better documents from your conversations. You want answers to your questions.

So we built Hyper. It's an iOS app where you tap to record any conversation (a 1:1 with your cofounder, a coffee chat, a standup, a brainstorming session). Then Hyper transcribes in real-time, and you can ask it questions mid-conversation ("hey hyper, what did we decide about pricing last time?") without breaking flow. Afterward, it generates notes and gives you one tap follow-ups (email drafts, slack messages) that integrate natively on your phone.

We're starting with this, but over the longer-term, we'd like to build a memory layer across every important conversation you've ever had. No files or folders or manual organization, just ask and get the answer. We're building towards a system that can resolve conflicting information, surface stale decisions, and untangle the messiest problems your team faces.

We're currently two founders, self-funded, building in SF. The design is intentionally bold and a bit weird. We wanted something that feels a bit more fun that traditional enterprise software, and ultimately something we ourselves would fall in love with. We'd love for you to try it in your next real conversation and tell us what you think.

— Shalin & Kanyes

As someone who's spent afternoons debugging a conversation I half-remember from standup, the idea of searchable real-world transcripts feels like replacing a whiteboard with a time machine. Curious how you handle the inevitable "I didn't actually mean that" moments when colleagues request deletions.

Heyy Shalin, nice launch! I've been waiting for something like this for ages. Out of curiosity, how are you approaching the consent issue? Are users expected to ask everyone if they could record them, does the platform offer any support for this issue?

One question: how does Hyper handle sensitive conversations? For 1:1s with cofounders or investor calls, some people will want certain things off the record. Is there a quick way to pause recording without it feeling disruptive? Congrats on the launch!

Capturing the conversation isn't the hard part — it's retrieval without noise. An hour of meeting has maybe 3 minutes of actual signal: the decision, the unresolved tension, the thing that got glossed over. Curious how Hyper handles the signal-to-noise problem — does it surface by time, by participant, or by something semantic?

Super interesting, whats different between you and other eKYC?

Since deepidv claims no third-party APIs, how do you handle global ID verification across 190+ countries. do you maintain your own document models and datasets for each region, or use a unified model that generalizes across different ID formats?

Good luck with the launch!

Congrats @shawnmarcmelo I’m curious about the economics at scale — is deepidv priced per check, per seat, or usage-tiered? For early-stage startups that need KYC compliance but aren’t yet doing high verification volumes, what does the entry point look like?

Running multiple parallel from one dashboard sounds like a huge sanity saver 😅 Terminal tabs get messy fast.

Congrats on the launch? ! This hits a very real nerve. As a founder running parallel product builds, the moment you start using more than one AI coding agent, you spend more time managing the agents than actually shipping. Quick question: Does King Mode let you set guardrails — like “don’t touch the database schema” or “only modify files in /frontend”?

Congrats @iritec_jp This is such a game changer! Does King Mode handle task dependencies intelligently? E.g., if Task 3 fails, does it block downstream tasks or attempt a workaround before flagging you?

As a non-techie, I have a question: How does kingcoding coordinate tasks between Claude, Codex, and Cursor simultaneously, is there a central orchestration layer that handles task decomposition, context sharing, and conflict resolution between agents?

Sounds interesting though, good luck!

Hey @daovid , I like the idea.
Wanted to try it out, but at the 3rd step of the creation i'm getting this error "That didn't work, User not found"