PH热榜 | 2026-04-28

Hey Product Hunters! 👋 this is Alex and I started Clera with @sebastian_scott3 and @daniel_wintermeyer.

With everything being built right now, there has never been a better time to work at a startup. But if you're looking for the right team and environment, you won't find it by just "applying" to roles. Clera is an AI talent agent that talks to candidates and hiring teams to build the kind of deep understanding that makes great matches possible. Then it introduces you directly.

Also introducing our $3m pre-seed fundraise. Find more info on LinkedIn and X.

Here's how it works:

• Tell Clera what you're looking for: your ideal role, dealbreakers, career goals. Via WhatsApp, iMessage, email, or our website.

• Clera's matching engine finds the best fit across 600+ startup jobs from VC-backed companies we work with directly.

• When you like a role, Clera makes the intro. If the company's excited too, you're connected straight to the hiring manager.

• Track everything in one dashboard: intros sent, interviews booked, feedback received.

What makes us different:

• Unlike job boards, we only list roles where we have direct relationships with the hiring team, so every match comes with a warm intro.

• We never force you into AI interviews. Clera is built to provide the best experience for top candidates, not to screen you out.

• Clera is completely free for candidates. Companies get started for free too and only pay when they make a hire.

We'd love to hear from you: Would you actually use this? What are the biggest pain points you run into when looking for jobs or talent?

Looking forward to your thoughts!

Alex

Nice product, slick UI and like the fact you can use iMessage / WhatsApp in the onboarding. Curious what is your vision longer term (post 5 years) for the future of traditional jobs and how do you see Clera adapting to that?

The conversational approach over iMessage/WhatsApp is a smart distribution choice. As someone who hired 100+ engineers over a few years as CTO, I can tell you the biggest matching problem isn't finding resumes that look right on paper - it's understanding what a candidate actually cares about vs. what they write in their LinkedIn headline. The dealbreakers and career goals part is key because those are the things that blow up offers after 3 rounds of interviews. If Clera can surface those mismatches early, that saves everyone weeks. Curious how you handle the signal quality on the company side - do hiring teams go through a similar conversational intake so the matching is genuinely two-sided, or is it primarily driven by job descriptions?

Looks like we have Bumble for job search. I gave it a try, looks pretty solid. Congrats on the launch guys :)

We've been using Clera at Pluno and already hired 2 engineers who came from Clera!
Quality of the profiles we receive is far higher than what we've had with traditional recruiters, and it saves us lots of time not having to manually reach out to great candidates.
Highly recommend, you guys rock @alexanderfarr @sebastian_scott3!

Congrats on the launch @alexanderfarr and the whole team!

How does Clera differ from platforms like Mercor, Braintrust, or even something like А Team for matching candidates with VC-backed startups?

love that this runs through iMessage/WhatsApp - feels way more natural than logging into another job board every day!

Literally direct intros to 600+ of the best startups in the US and Europe. If I weren't working for them, I'd be the first one chatting with Clera to get intros.

This looks incredibly promising. I really like the conversational approach, it feels like a much more natural and engaging ways to discover roles. Congrats on the launch 🙌

Now is the time to build agentic solutions that will make our lives easier. Building a recruiting agent that changes the industry in such a positive way is one of the most exciting things I have ever done.

Been using this and honestly… this is how job hunting should work. Super clean matchmaking and actually relevant intros. Congrats on the launch guys!!

Building Clera has been the joy of my life. Shipping product that actually help people land a job is a dream come true! I have seen countless friends / family stuck in jobs they don't enjoy. With Clera we can finally remove that friction!

Keen to learn about any feedback while using the product!

Crazy product (ok, I am biased). Happy launch day, go Clera, go!

Smooth platform from the candidate's side :)

clera found our first gtm hire!

Congrats guys! Great team! On a side note: How do you find and match candidates?

running through iMessage is a smart call. job searching already feels like a second job — the last thing you want is another app to check. curious how it gets to know your preferences over time, like does it ask directly or pick up on what you react to?

Congratss 🎉 Love that the bar is "intros where the company is actually excited" rather than another firehose of listings. The WhatsApp/iMessage entry point is genius!!!

a love story but with stock options. (forever, but four years). lfg!!

I had a great experience using Clera during my recent job search. The platform’s ability to handle company-specific 'additional questions' allowed me to provide high-signal responses that actually showcased my background. It clearly makes a difference—I was able to secure an interview with one of the companies that I see great match. Highly recommended for anyone looking for a more tailored recruiting experience.

Nice. Love the idea and the video.

Are you planning to find more CTOs to become Member of Technical Staff ? :)

https://x.com/henrythe9ths/status/2049148130059292743?s=46&t=3M0KdWHI2R4FjKbL-WWA5A

Congrats and good luck!

Looks really cool - can't wait to try when we restart hiring. Curious how do you source candidates - is it Clera users on the other side; or also off-platform candidates?

Big fan! Love how easy it is to get started and meeting all the great talent!

i am assuming its primarily focusing on the west for now?

Super smooth onboarding, found 3 solid SF roles that actually match my background. Love it!

Sounds really interesting! How do you make the intros after I've shown interest in a company? Also curious if Clera helps tailor the pitch for me before the intro, or if it's a direct connection to the hiring manager?

Feels like the hardest part here isn’t finding roles, but actually understanding what someone truly wants (which most people struggle to articulate).

Curious how Clera gets signal beyond what users explicitly say?

We have been working with Clera for months now, and it has become a key hiring channel for us.

with whatsapp - right top of mind, smart!

Love the "CEO not debugger" line.
Real question though: when the COO, CMO, and CTO agents share one memory, who wins when there's a conflict between sales goals and product reality? 😅

BTW, many congrats on the launch team!

@SureThing.io best agent I've ever used so far. I'm using it for my linkedin post ideation to posting. how about adding personal skills?

I have like 30 starred repos and maybe 2 running lol. Congrats on shipping this, curious how the memory layer works across the COO/CMO/CTO agents — is it shared context or do they each have separate threads that sync?

@celine_yu Congrats on the launch

Quick question: if agents report up like humans, does that mean I'll start getting "quick syncs" and "status update requests" from my AI at 5 PM on a Friday? 😄

Love the chairman, not coder framing. How much business context can SureThing actually absorb before it starts making decisions?

Building SureThing is fun. Seeing how people put it to work is what makes the late nights worth it.

A few snapshots users have shared with us:

→ A B2B founder who finally turned LinkedIn into a real pipeline. SureThing finds the right people, opens the conversation, follows up for weeks, and hands him warm leads when it's actually time to talk.

→ A therapist building her public practice on X and Instagram. Content drafted, posts scheduled, comments answered in her voice. The followers showed up faster than she'd planned for.

→ A Shopify seller spotting which products are about to pop, generating the listings, refining the copy, and watching her conversion rate climb.

→ My favorite: a mom who built a little math game for her kid. No code, no engineer in the family. Just one evening, one idea, and a lot of love. It's part of the bedtime routine now.

These aren't features we built. They're things people figured out to do with SureThing, and every time we see one, we want to make it possible for more.

Make your goals a sure thing.

Congrats! Quick q — when you paste a GitHub skill, how much manual config is still needed to get it running with business context? Is that part guided or does it require some technical setup upfront?

Does SureThing work with private repos or just public starred ones?

What happens when the AI makes a bad call is there an undo or audit log built into the dashboard?

“Agents that work like a team” sounds great,
but getting consistent results from even one AI agent is still tricky.

This hits.

Feels like the real problem isn’t building more agents, it’s getting them to actually work together and move things forward.

The “agents that report up” idea is interesting — closer to how people actually think about running a team.

Curious how the shared memory holds up in practice. That’s probably where it either clicks or breaks.

Congratulations on the launch! If SureThing “does the job” instead of just suggesting, where do you draw the line between autonomy and human oversight?

Looking forward to see how SureThing develops

Very interesting product. quick question: the agent can get certain skills from GitHub, how would the agent be suitable for a certain job in that specific startup. Could the agent be connected with working software like emails or slack so that the agent has all context?

My first reaction: this feels less like "another AI agent" and more like the missing layer between starring a repo and actually running it.

@agentos Could you share more about workforce scaling and latency? I’ve run into systems like this before, and in production environments these issues usually become more significant.

Congrats on the launch, is it fully autonomys and is there a possibility to add human input to the loop, in order to make sure it's compliant with business goals?

If it's done, how do you handle it?

the "chairman not debugger" framing hits differently when you're a solo founder wearing all three hats. COO, CMO, CTO — same person, same head. agents that report up instead of waiting to be queried is the actual shift.

his sounds like a dream. As a solo founder working warehouse shifts, I spend more time being a 'debugger' on my Chromebook than actually being a CEO. My biggest bottleneck isn't the AI code, it's the 'human' energy to manage all the silos (CMO, CTO, COO) alone. How does SureThing handle the deployment side of things when the 'CTO agent' hits a wall with infrastructure limits?

Exactly the problem. Open-source AI tools are amazing, but the setup overhead kills adoption. Nice solution.

This resonates a lot — tons of starred repos, almost none actually used.

If SureThing can bridge that gap reliably, that’s huge.

The “starred vs actually running” gap is real. Most tools like OpenClaw or Claude Code still assume users operate in a terminal. SureThing.io's reporting line framing is a clear shift toward outcomes, not setup.

this is exactly what non-technical founders have been missing from the agent wave. congrats, following closely 👀

Wow, I like the feature of pasting any GitHub skill, which will largely accelerate the developer's workflow. Nice done!!!

The "Karpathy research agent / Garry Tan gstack" framing nails the real friction - discoverability is fine, but most starred repos rot, and "right there" usually means broken installs. Curious about the QA layer. When you turn a GitHub skill into a teammate, what catches dead repos vs. the working ones - runtime tests, community signal, or trust the maintainer? And once a skill drifts (a dep breaks, an API moves), does the employee silently fail or know to escalate?

goodluck on your launch Celine! Could i interact with SureThings via whatsapp?

SureThing is an impressive innovation in AI agency solutions! By transforming GitHub skills into a seamless, collaborative team, it effectively breaks down silos within organizations. The persistent memory feature across key roles ensures that everyone is aligned and working towards shared goals, allowing for a more agile CEO-like management style. This is especially crucial in today’s fast-paced business environment where efficiency can make or break a company. If you’re looking to elevate your business strategy and enhance team collaboration, SureThing might be just what you need.

Love this tool! Need to try it asap.

the 'reporting line vs terminal' framing is sharp. the hard part nobody's nailed yet is getting the agent to be honest about what it almost-did, especially silent partials like 'i created the event' when the api returned 403. how do you surface that in the dashboard?

👋 Hey Product Hunt fam!
I’m Luke, and I’m excited to launch Social Fetch today.

🌐 What’s Social Fetch?
Social Fetch is a single REST API for real-time public social data—profiles, posts, comments, videos, transcripts, and metrics, across TikTok, Instagram, Twitter, and more. You get predictable JSON and one integration path, instead of a pile of one-off scrapers that break every time a layout changes.

💡 Why I built it
Every product that touches social eventually reinvents the same nightmare: bespoke parsers, surprise schema changes, and late-night incidents because a selector moved. I wanted social data to feel boring in the good way—stable fields, documented errors, and a requestId on every call so support can actually help. Your team ships features; the API absorbs the churn.

🚀 Why try it?

• ⚡ One API for many networks—add coverage without rewriting your stack

• 🧱 Normalized responses—same concepts across platforms, fewer edge cases in your code

• 🤖 Built for builders—OpenAPI, official TypeScript SDK, and llms.txt / llms.json so humans and coding agents can integrate fast

• 🛠️ Fits how you already work—backend services, internal tools, analytics pipelines, creator workflows

• 🔑 Debuggable by design—when something’s off, you’re not guessing; you have a request ID and a clear error envelope

🙌 Join me
I’d love you to sign up, hit the playground, and tell me what you’re building. Comments, feature ideas, and upvotes all help prioritize what ships next.

Start free on the site, or reach out anytime: 📨 support@socialfetch.dev

Congratulations on the launch. @lukem121

why am i seeing so many relevant products lately, thank you for building this- i am curious how are you finding your way around these social medias blocking scrappers? are you integrating with other permitted tools? would love to get a take on the consistency of results.
again great product luke

Unified APIs for social data always look great…
stability and consistency across platforms is usually the hard part.

Interesting product, but how does it really work on the backend? Do you scrape social media platforms?

One API across six platforms is genuinely hard - the cat-and-mouse on LinkedIn alone usually wrecks people. The 'no rate limits' line is the part that makes me curious. Is that no client-side rate limit (you're absorbing the upstream cost) or are you proxying through a residential pool that quietly degrades at scale? And how do you handle platform-level breakage - when LinkedIn ships a new auth gate next quarter, what's the realistic time-to-fix on your end?

Tried building an all-in-one influencer resume tool a while back — wish I'd had this then. Would've saved me weeks of scraping headaches!

WOOOOOOOOOOOOOOOW!!!!! The possibilities I can do with this and the help with Ai. So much times I've wanted to do research, gather data from social media platforms, but I come against either platforms that are just WAY TOO EXPENSIVE (the good ones are super expensive) or to cheap and they don't have the features I need. But now from reading the home page, this is what I've been waiting for!

Amazing!!!! @lukem121

is the website inspo from wispflow?

The llms.txt / llms.json support is a detail i haven't seen other data APIs do yet. makes sense given how many people are piping social data into agents now. Curious how you handle linkedin — that one always seems to be the flakiest. is coverage there stable or still hit or miss?

When will you improve your code generation quality? This is the part that really needs significant improvement.

Boring designs, rigid Vibe coding templates and high credit consumption led me to switch from the Pro Plan to exploring alternatives like Claude.

I appreciate why some people want an app but I'm happy with the web version so not sure why I'd switch.

I'm very excited about the native mobile launch of Lovable. I have been using Lovable since June of 2025 to build my product (web, iOS, and Android). I have also built all my marketing pages with Lovable. Fantastic experience. So much flexibility. With imagination and critical thinking (and curiosity) you can build anything. The native mobile app makes that easier on the go.

I like the app's design but after a few experiments I'm still not sure who the ideal user is. My developer background might be creating a bias — this could simply be a tool that resonates more with non-technical users.

Hey Product Hunt 👋 - I'm Tahiya. We spent years watching AI teams hit the same wall: the moment they tried to move their applications outside the cloud - to a factory floor, an edge device - their vector database stopped working. Latency spiked, connectivity dropped, data residency requirements kicked in. The infrastructure just wasn't built for it.

We've seen that most vector databases were designed for the cloud, and that was fine when AI lived there. But AI doesn't anymore. It's moving to edge devices, disconnected field environments, and embedded systems. And cloud-based databases break the moment you leave the data center.

Actian VectorAI DB is a portable vector database built for exactly this reality. You can run it on a Raspberry Pi, an NVIDIA Jetson, on-prem behind a firewall, or in the cloud - using the exact same API and architecture throughout. No re-platforming. No re-architecting.

We're launching GA today. In VectorDBBench tests at 10M vectors on identical self-hosted hardware - with zero vendor optimizations applied to any database - VectorAI DB delivered a 22x QPS advantage over Milvus and Qdrant, retaining 72% of its throughput at scale while competitors dropped to ~12% of theirs.

You can build on VectorAI DB today for:
• RAG pipelines (local, edge, or hybrid)
• Monitoring & anomaly detection
• Enterprise semantic search

Python and JavaScript SDKs. LangChain, LlamaIndex, and Hugging Face support. Runs as a Docker container: Kubernetes, Helm and Terraform compatible. Linux and Windows are supported, both on ARM and x86. Compliance-ready for ISO 27001, SOC 2 Type II, HIPAA, and GDPR.

We're building for teams who can't compromise on where their data lives. If that's you - grab the community edition or free trial, join us on Discord, and tell us what you're working on. We're reading every comment today. 🙏

I'm always a big fan of on-prem/local support. Congrats on the launch!

portable vector db is exactly what's missing in this space. most solutions lock you into their cloud infrastructure which kills flexibility. what's the memory footprint like for embedded deployments? thinking about IoT scenarios where you're super constrained on resources.

Great work, congrats on the launch! :)

Super cool, congrats on the launch!

Looks great!

Edge vector DBs are a real underserved space - most "portable" claims fall apart the second you actually deploy on a Jetson or factory PC. Curious about two things on the 22x QPS claim. What is the storage footprint at 10M vectors vs Milvus on the same hardware - is the speedup paid for in disk? And on hybrid deployments, when an edge index syncs back to a cloud cluster, is it a full rebuild or do indexes merge incrementally?

curious how it handles intermittent connectivity — like if an edge device goes offline mid-query, does it fail gracefully or does it need a persistent connection to work?

interesting to see focus on edge deployment. we've been running into latency issues with cloud vector searches for real-time wearable data processing. how does the performance hold up when you're doing frequent updates to the embeddings, not just reads? the 22x claim is impressive but curious about write performance.

🦁 Hello Product Hunt community!

I'm Alex, Co-Founder & CRO @ Crono. This is our fourth launch, and yes, we keep coming back stronger every single time.

Launch #1: underdogs. We won. 🥇
Launch #2: we came back stronger.
Launch #3: we introduced Crono 2.0, the AI GTM Platform.
Launch #4: we're not launching a feature. We're launching a new era of B2B sales.

⚡ Introducing The Agentic Sales Engine.

After working with 300+ sales teams across Europe, we kept seeing the same thing: the best reps weren't losing on skill, they were losing on execution. Too many tools. Too much manual work. Too much noise, too little signal.

❌ The problem isn't effort. It's architecture:
↳ Sales reps spend ~70% of their time on non-selling activities
↳ Only 28% of B2B sales teams hit quota in 2025
↳ Autonomous agents that act alone lose context and make things worse

✅ So we rebuilt the execution layer from scratch.

🚀 Crono is now the place where humans and AI agents work side by side: → Real-time signals captured automatically (job changes, hiring activity, website visits, engagement) → Context built for every account, so reps always know why and when to reach out → Workflows that execute themselves: prospecting, enrichment, outreach, follow-ups → Agents handle the repetitive. Humans close the deals.

Most sales tools help you manage pipeline. Crono helps you execute it.

Data becomes signals. Signals become context. Context becomes execution. Execution drives revenue.

Already at 1M€ ARR, growing 4x YoY, and just getting started. Trusted by Bizaway, Alibaba, Factorial, Busuu and 300+ teams worldwide.

Works with HubSpot, Salesforce, Pipedrive, Gmail, LinkedIn, Clay, n8n and more, without disrupting what already works.

We're here all day, ask us anything, break our demo, challenge our thinking. That's what PH is for. 👋

⚡ The agentic sales revolution starts now. Let's go!!!

Hey Alex! Signal-driven prioritization is a promise many tools make: what are the concrete signals Crono has found to be most predictive (job change, hiring, web intent, engagement, etc.), and how do you prevent reps from getting “faster noise” when signals conflict or are low-quality?

Hey PH 👋

We started Famnest because family life today is scattered across too many places. Messages in one app, schedules in another, important health info buried in notes, and bills somewhere else entirely. It works… until something slips.

So we asked ourselves: what would it look like if everything a family needs actually lived in one place?

Here’s how we’re using it ourselves.

Our family calendar isn’t just dates, it’s birthdays, school events, and the small everyday logistics that keep things running. Important info like allergies, contacts, and notes are structured so nothing gets lost or forgotten. Bills and shared expenses are visible without awkward reminders or chasing.

One thing that’s made a bigger difference than we expected is the daily digest. Instead of checking multiple apps or asking around, you get a clear overview of what’s happening today, what’s coming up, and anything that needs attention. It sounds simple, but it removes a lot of small daily friction.

The biggest shift for us has been coordination. Fewer messages, fewer repeated conversations, and far less “who’s handling this?”. Everything is already there, shared, and clear.

We’re building this with privacy at the core. Family data is deeply personal, and we think it should stay that way. No selling data, no unnecessary exposure, just a secure space for the people who matter most.

We’re still early, and we’re shaping this together with the people using it. We’ve also started a small (and growing) subreddit where people share how they organize family life and what they wish existed, feel free to join us there as well.

Question for PH: what’s the one thing in your family life that always falls through the cracks?

Would genuinely love to hear how others are dealing with this.

I love Famnest! As someone who has to coordinate a ton of multiple family medical appointments and keep medical information for each it’s been a lifesaver. Everything is in one app! I am also very impressed and happy that they take user suggestions seriously. To me it doesn’t feel like a user/developer relationship. It feels more like family members discussing a product. In today’s world where everyone is behind a screen it’s almost impossible to get that sense of working with a close friend or family member and not a stranger. Famnest is different and I would definitely recommend this app to anyone.

this is super useful for young families and larger families as well. I can see how we can share responsibilities across siblings as well with aging parents and niece/nephew care! Super cool, definitely gonna suggest this in the family group chat

Hi! Personally since I have found this app it has been so helpful in my everyday life! Nothing scattered anymore, days all organised. I love it and don’t know what I would do to organise my day without it!

Hi again 👋

We're back on Product Hunt, and this time with a feature that I'm most excited to talk about. AI Reports.

Honest confession. Writing reports sucks. I hate it with all my heart.

Running the analysis and putting together a dashboard with cool charts and clever insights is the fun part, but turning all of that into a report you can defend in front of leadership is where it stops being fun.

For me, dashboards show you what happened. But a report shows you what someone close to the work thinks about what happened. That used to take weeks, but now it doesn't.

That's what AI Reports does, and it's one of my favourite things @yash_gandhi3, @navdeep710 and I have shipped. You run your analysis in @OrcaSheets, give it a template for how your team writes things up, and it generates a fresh report on the new data - in minutes.

Here's the thing though. AI gets lost in millions of rows of raw data. But give it a few hundred rows of aggregated data and a template to work against, and it does the job honestly well.

AI Reports is normally a paid feature, but if you're coming in through PH, it's free until next Tuesday 💥 Try it on a report you're already dreading. I'll be here all day to answer any and all questions.

Tell me: what's the report you write every month that you wish you could just generate?

Dashboards do one job really well, which is showing you what happened without anyone's opinion attached. That's exactly why they have a ceiling on how useful they can be.

A report does the opposite job. It's written from inside the work, by someone close enough to a project to hold the numbers against what they actually know about how things went. The output isn't what happened, it's what someone responsible for the work thinks about what happened.

The hard part of building AI Reports was figuring out how to keep that perspective in the loop without keeping a human's time as the bottleneck. The answer turned out to be templates. A team writes a template that captures how they think and frame their work, the AI runs against that template on new data, and the separation is what makes the output reliable in a way prompting an AI directly on raw data never quite is.

Genuine ques - where's the actual bottleneck when you write a report? The analysis or the writing?

Monday blues needed a rescue. Here it is :)

Hey PH. I am Nazz - the Creator 👋.

I built WUPHF for myself at Nex. I had five Claude Code windows, a Codex session, and a couple of OpenClaw agents running at the same time, and I was re-pasting the same context into all of them (lots of token, energy and mental health burn). When I wanted them to coordinate, I was the manual relay. Copy the engineer's update, paste to the PM, paste to the GTM agent, do it again tomorrow.

I tried the multi-agent frameworks already out there. Every one was some flavor of Paperclip with a Linear-style DAG on top. Write a plan, watch nodes turn green. Functional, but I did not want a project-management dashboard. I wanted the interface I already use to get work done with humans. Channels, @mentions, DMs, threads. I wanted to chat with my agents the way I chat with my team.

So I built that over a weekend. Used it for a week. Realized I was not going back. Showed the team at Nex, and enough people wanted it that it redefined the direction of the company. WUPHF is now Nex's open source product.

The shape: collaborative office for AI agents. Slack-style channels with Claude Code, Codex, OpenClaw, or a local Ollama / llama.cpp endpoint via OpenCode as the members. They learn and run your playbooks 24x7.

Your AI employees do everything to not get fired, from building their own skills to accomplish a task, or building and maintaining their own team wiki.

The team wiki is a Karpathy-style LLM wiki on markdown + git in ~/.wuphf/wiki/. Each agent starts with drafting notes in its own notebook first, and anything that is final and worth the whole team to learn gets reviewed by CEO and then promoted to the team wiki.

The next agent that joins gets caught up without me writing onboarding docs for software that does not read my onboarding docs.

Open source (MIT license), self-hosted, your keys. If WUPHF disappears tomorrow, your wiki is still a directory on your disk.

Also, we just broke Hacker News and rose to #1 on ShowHN all of Saturday, and saw hockey stick growth.

Install: npx wuphf@latest
Repo: https://github.com/nex-crm/wuphf
Website & Demo video: https://wuphf.team

I wish slack generated wikis out of my conversations like this!

It's been fun experimenting with the versatility of the agent team. Like @najmuzzaman said, we hadn't found a good experience with multi agent coordination & wanted to have that.

It's also been the best use of my local LLM setup. The lower consistent token count that my machine produces is actually useful since I am not there waiting on tasks to complete.

already using it. very cool. love the style and the autonomy of each agent!

for the uninitiated.... what does WUPHE stand for??? 🫣

smart that the wiki is just markdown + git on disk. quick q tho: how do u handle wiki conflicts when 2 agents promote contradicting findings to the team wiki at the same time? CEO agent arbitrates or last write wins?

Hey Product Hunters! 👋

I'm Krzysztof, creator of SimCam at Software Mansion – and I'm super excited to finally share this with you! 🚀

I built SimCam to solve a common frustration: the iOS simulator simply doesn't support camera testing.

🎥 With SimCam you can:

• Use your Mac's camera as the "iPhone camera" in your apps running on Simulator

• Inject any image or video as a camera feed

• Generate QR codes and inject them into the stream

• Control the front and back cameras independently

• Control iOS simulator camera source programmatically via CLI or let AI agent use it to build and test camera scenarios in your app

• All that works without adding any dependencies or you changing anything in your app

With SimCam you no longer need to test your apps on device only to verify camera enabled features!

It’s available as a one-time purchase with a lifetime license. Give it a try at simcam.swmansion.com.

I'd love to hear what you think – any feedback is super welcome! 🙌

MaxHermes is a cloud sandbox AI agent from MiniMax that autonomously extracts reusable skills from completed tasks and applies them in future sessions.

Problem → Solution: The statelesness of AI agents is a design problem that the industry has mostly ignored. You do the work of teaching the agent your context, it performs, the session ends, and nothing carries forward.

MaxHermes introduces a learning loop at the execution level. Every complex task triggers automatic skill extraction. Those skill documents persist, get refined through feedback, and stack over time. The agent that handles your tasks in month three is meaningfully more capable than the one you started with.

What makes it different: Built on MiniMax M2.7, it achieves this without requiring any manual capability setup or server configuration. The evolution is autonomous.

Key features:

• Learning loop that auto-generates skills post-task

• Cross-session persistent memory

• 7x24 cloud availability with no infrastructure on your end

• Integrations with Feishu, DingTalk, and Enterprise WeChat out of the box

Benefits:

• Your AI assistant improves from real usage, not from prompt tuning

• Zero setup cost for enterprise teams already in Chinese productivity ecosystems

• Parallel task execution means it handles multiple workstreams without queuing

The compounding skill model is the bet worth watching here. Whether it holds up at scale and across diverse task types is the open question at launch.

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

Congratulations

cloud sandbox approach makes sense for this. how does the skill library work across team members? wondering if there's a way to share learned skills between agents or if each instance builds its own knowledge base.

Hey! 👋

I'm Luis, solo dev behind Curflow.

If you've ever used mouse gestures in your browser (right-click + drag to close a tab, go back, open new tab) — Curflow does that across the entire system. Finder, Xcode, Slack, anywhere.

For trackpad users: macOS gestures are great, but you're limited to what Apple decided. Curflow lets you create your own. Draw a gesture with a modifier key and trigger any action.

For mouse users: you lose all the trackpad stuff. Mission Control, swiping between desktops, gesture navigation. Curflow basically turns any mouse into the Apple Magic Mouse we all wish existed — full gesture control, no compromise.

One thing I didn't expect: people are telling me it's useful for accessibility too — when you only have one hand free (holding a baby, coffee, or permanent/temporary disability), gestures let you do more without reaching for the keyboard.

 
14-day free trial. I'm here all day!

I do not think that there are so many solutions like this. So it is even more unique :)

As a fellow maker currently building in the Ai Image space, I’m wondering if we could use this as a tool to give our customers a little more control over their templates. How easily does it plug into other work flows?
We've built in Mind Studio but are also going 'headless'

Very cool - love products like this that don't try to do too much but nail simple functionality. Clicked the link and played around with it for half an hour without even realizing

Now this is a pretty interesting and unique product! It's kinda refreshing in this day and age of sloppy content to see a nice & well thought tool to create interesting visual elements and this is something I would definitely buy!

But it's probably where my issue lies: I would love to buy a downloadable version of this product even if I have to pay 100$ (so more than the annual fee) to use as a local app but I don't see myself paying a subscription to generate gradients, no matter how interesting it is, as I could do this with Illustrator in some ways.

I think you have a very interesting tool, but I'm not sure you got the appropriate business model. I know everyone want to have their monthly revenue and therefore go for the subscription model, but it's not appropriate for all businesses, and some products are better suited for one-time purchase with yearly updates that you can pay for if you choose to (Xnapper is a real good example of this and got my purchase immediately).

I wish you the best, and if you decide to offer an alternative way of purchasing your product for local usage without requiring subscription, you already got a customer for that!

All the best buddy 🙌

I tried the app and it's really cool. For the future - how will it handle animations? What export options will you support?

Hey Product Hunt 👋

We built Voice Agents for MindPal because a lot of expert-based businesses have the same bottleneck:

their best knowledge is trapped in calls, courses, PDFs, Slack threads, and repeated client conversations.

MindPal already helps coaches, consultants, agencies, educators, and creators turn their expertise into shareable AI agents and workflows. But for many client-facing use cases, chat still feels one step removed from the way people naturally ask for help.

So we added voice.

Use cases:

• 24/7 client support

• Content creation by talking

• Sales roleplay and objection practice

• Course/community support

• Website or client-portal assistants

We would love your feedback:

- What voice-agent use case would you build first?

- Would you use this more for clients, team training, or your audience/community?

- What should we improve before this becomes part of your daily workflow?

Happy to answer questions all day. Thank you for checking it out 🙏

Nayana, it was wonderful iterating to simplify and refine the IDE plugin together, congrats on the launch!

when you're working solo, there's no senior engineer to push back on a vague task. you're both the person writing the prompt and the person who should have asked better questions before writing it. ide extensions make sense — wondering if jetbrains / webstorm support is on the roadmap, or is the vs code ecosystem the focus for now.

free, open source, and available in the editors people are already in rather than a new tool to learn is the right distribution bet for developer tooling. adoption usually dies at the signup screen.

This nails a problem I saw constantly as CTO scaling an engineering team from 15 to 120 people. The gap between "what the developer typed into the prompt" and "what they actually needed" was the single biggest source of wasted agent cycles. Junior engineers especially would fire off a vague prompt, get back 200 lines of plausible-looking code, and spend hours debugging something that was wrong from the first instruction. The insight that the back-and-forth questioning phase is more valuable than the final plan really resonates - it mirrors how the best senior engineers work. They spend 80% of their time clarifying the problem and 20% solving it. Blueprint essentially automates the "senior engineer asking the right questions" step. Curious whether you're seeing different question patterns emerge across codebases - like whether Blueprint asks fundamentally different clarifying questions for a microservices repo vs. a monolith.

Hey PH 👋 I'm Yota, one of the makers of Jitera.

Turn AI into your teammate. With shared context.

Over the past year, AI went from novelty to daily tool. But something still feels off.

Everyone prompts their own AI in their own tab. Outputs get copy-pasted, nobody reads them, and the team falls out of sync. Your organization isn't learning, it's just generating.

Even with AI everywhere, teamwork itself hasn't changed.

Here's what we believe: without context, AI is a genius goldfish, brilliant in the moment, then forgets everything. With shared context, it becomes your teammate.

Jitera builds that shared context as a Context Graph — connecting your code, docs, decisions, and tribal knowledge so every agent knows who owns what, what was decided, and what's already been tried.

That's Jitera.

What it looks like 👇

🧭 Context Graph
Connect docs, data, decisions, and people into shared context your agents can actually use.

🌱 Agents that grow your context
Agents ask the right teammate when context is missing — so your shared context keeps getting sharper.

📝 Humans and agents co-edit
Docs, specs, and notes edited together by humans and AI, in the same document.

💬 One shared thread, not 100 private tabs
The whole team chats with agents in one place. No more copy-paste from private ChatGPT windows.

Not just faster outputs. Better decisions, made together.

We're early, and we're building this with teams who want AI that actually fits how they work. If that's you, we'd love your feedback — the sharp kind 🙏

👉 https://jitera.com/

— Yota & the Jitera team

Hey, Ivan here, I work on Jitera.

Before this I spent a few years building open-source tooling for people running their own LLMs and agent stacks, so I came into the project with strong opinions about what agent platforms tend to get wrong.

I want to talk about the part of Jitera I'm very proud of, which isn't really something you'd put on a marketing page. It's how the platform is built underneath.

The obvious way to build something like this is to make one big core that handles everything itself. The platform decides how memory works, where telemetry goes, what file storage looks like, which LLM runtime gets used. That demos great, but breaks down when someone asks if their agents can route usage data to their own observability stack instead of ours, and another wants their agents to mount an S3 bucket as if it were a local filesystem, and a third wants to swap the LLM runtime to Anthropic's Claude SDK for one specific team's agents but not the rest. In the all-in-one version, each of those becomes its own fork in the codebase, and a year later it transforms into something nobody can reason about.

We built it differently, so every behavior in the platform is a small piece of middleware wrapped around the agent loop. Memory, telemetry, input classification, the cloud bucket mounts, even the LLM runtime itself, they all have the same shape and they stack on top of each other without interfering. Adding "always preload this URL into the agent's context" is just a tiny self-contained module, and the bucket mount and the per-agent telemetry sink are around the same. New behaviors plug in without making the platform more complicated. One interesting bit is that Jitera's own agents are built on the same exact primitives, so anyone using our products can replicate or enhance our agents on their own.

I've worked on enough infrastructure to know this is the kind of decision that pays you back every time and really proud of everything this direction enabled us to do.

I really hope that you'll like the product's flexibility and features, excited to see how it'll be used :)

Jitera team member here 👋

One thing I personally feel more and more in daily work: even with AI everywhere, great products are still built by teams.

AI can generate fast outputs, but real product development still needs discussion, shared understanding, feedback loops, and people aligning together. That’s why I’m excited about what we’re building at Jitera — not AI replacing the team, but humans and agents working together as teammates.

What makes this especially interesting to me is the “team in the loop” part. As agents interact with people across the team, they gradually understand more of the company’s tribal knowledge — past decisions, context behind discussions, ownership, constraints, and how the team actually works.

The context grows with the team, instead of staying trapped in isolated chats and private tabs.

Really proud to be part of a team building toward that direction 🙏

A shared context graph for agents is super interesting.

Curious: can agents adapt to individual working styles too (how someone writes, makes decisions, reviews), or is it mostly built around shared team context?

Hey PH, I'm Nao, founder of Jitera!

Without context, AI is a genius goldfish. With context, your teammate.

Jitera has three layers — Documents, Memory, and AI Agents.

By giving each agent its own context, we designed AI that works like a real teammate.

In Japan, we've been the context layer for multiple Fortune 500 and global companies — especially powerful for reverse engineering and product management in software development. Today, we're launching in the US!

Hey PH 👋 Keima here, one of the makers of Jitera.

Recently, AI agents have become incredibly capable. Tools like Claude Code, Codex, and OpenClaw can look almost autonomous. But when you compare them to how humans actually work, there’s still a noticeable gap in the quality of real output.

We think that gap comes from context. Jitera is built around a Context Graph that continuously builds and updates the context an agent needs to do meaningful work. Agents don’t just use context. They grow it. They act based on what the team knows, and at the same time, accumulate new context through their work. In a way, they behave almost like something that “lives” on information.

Because of this, when you use Jitera as a team, the context that used to live only inside individual team members gradually becomes shared with agents. The result is a new kind of team experience, where humans and agents operate with the same understanding.

It’s still early, but we’d really love for you to try it with your team and share your feedback 🔥

Hey ProductHunt! 👋 I'm Paula, the founder of Oly.

Six years ago I started working in fashion trying to solve the problem of excess inventory. I built my first startup around upcycling — helping brands give unsold items a second life. But I quickly realized upcycling was hard to scale, and meanwhile the resale market was exploding.

So I pivoted. A fashion brand hired me to sell their "not new" inventory across resale platforms and I spent weeks doing it manually — listing each item one by one across Vinted, Vestiaire, Grailed, each with completely different requirements. No tool existed for this. Every solution was built for new products and completely broke down on secondhand complexity.

So I built Oly.

This is my second company, and the one I was always meant to build. I bootstrapped the entire way — reinvesting every euro of profit back into the business, staying lean, staying focused. That discipline is what got us to six figures with two people and zero marketing spend.

Today Oly manages over 1 million product listings, has processed $2M GMV in Q1, and we're growing 25% month on month — with zero paid acquisition. Grailed, the leading men's resale marketplace in the US, approached us to power their seller integrations.

We're just getting started. Would love to hear from anyone building in resale, fashion or circular economy — and of course any feedback on the product!

Syncing listings across so many platforms sounds useful,
but I wonder how well it handles differences between each marketplace.

The best app!!! Don’t have to use them annoying pay and displays anymore or have to get out my car and wait in the rain trying to pay for my parking or have to rush back when I’m about to run out. Never have to think about parking again it literally does everything for you all you have to do is park and walk off, it’s going to save me so much money.

Just so much easier, you literally park your car and the app pays for your parking. It knows what bay you're in, it knows all the codes/ areas/ locations/ rules, I don’t have to think about it anymore. I have no clue how it works but I love it, it means I online one parking app and that’s Parc. I’ve told all my friends to use it. It’s just such a good idea (It’s also especially great for those who get a lot of parking fines).

Nice one!!- how did you manage to connect to the providers? Would love to see something similar in Austria & Germany!

The simplicity of this is genius. One absolute huge stress I never need to worry about any more

I’m constantly parking when I’m rushing around, school runs, appointments, food shops etc and I used to always forget to extend my parking. So I’ve had so many fines because I’d lose track of time. I literally just park and leave. It sorts everything automatically and I don’t have to panic about checking my phone.

There are already a bunch of parking apps, but most still require manual input and constant checking. What stands out here is the automation layer, from my pov its the auto start n auto stop and extending sessions to avoid fine! . That shift from tool to something that just quietly works in the background is powerful.

Hi Hunters! Alisa here, founder of Immersive Fox.

I built this after watching Fortune 500s buy Microsoft Copilot for 50,000 employees, then spend six months still preparing pitch decks manually. Tools without training are just expensive shelfware.

So we built AI agents that turn any document into a full multilingual training course in 10 minutes. SOPs in, courses out.

Why we are different in a crowded category:

Speed. 10 minutes from a PDF to a complete course. Category average: two weeks.

End-to-end loop. Articulate gives you a builder. Synthesia gives you AI video. We give you generation, delivery, comprehension, and on-the-job application in one system.

40+ languages, properly localized. Your French team gets a French course that reads like it was written in Paris.

Conversational AI tutor. Real-time Q&A on the content. Your learners ask, the tutor answers from your source material.

SCORM-ready. Works with your existing LMS. No migration.

What is the slowest piece of training you have sat through this year?

Alisa