PH热榜 | 2026-03-26

I got connected to the Littlebird team through @joshconstine. Their approach to the never-ending pursuit of obtaining more context for AI is bold...!

Unlike Instagram after 15 years, most AI tools struggle to know things about me. I do use some memory tools here and there, but until Apple revamps Siri, what has the most digital context about me (besides my phone)? My laptop, duh!

Which is why Littlebird "sits" on my shoulder, reading my screen (capturing only text, not screenshots) and listening to my meetings, so that when I ask it something, it already knows what I've been working on.

And before you ask: yes, you can exclude specific apps if you don't want Littlebird nosing (beaking??) into all your stuff. In fact, you control exactly which apps it can see, and can pause or delete any data it's collected. It’s SOC 2 certified, if that's important to you. And unlike GitHub, they never train on your data.

Give it a few days and ask it “what have I been working on this week?” Come back and tell us how much Littlebird knows!

my life has been easier since i started using littlebird. it runs on the side, watches on the side - and then one day you're mid-conversation and instead of scrambling you just go "let me ask littlebird." you stop losing things. that's basically it.

The "full context from your screen" framing is interesting - where does the screen data actually live? Local only, or does it go through your servers? That's usually the first thing people ask before they'll actually use something like this.

@dhaval_singh2 introduced me to LittleBird! The thing that stuck with me is I stopped explaining myself to it.

Every AI tool has that setup tax where you brief it before it helps you. Littlebird just already knows. I ask and it answers. No "here's what I'm working on."

One moment that got me was I asked it to review a technical plan and it caught a gap that would've broken things downstream. Without me explaining any context.

This shows that it's not just memory but it reasons about your work with that memory. That's the actual difference.

Congrats on the launch team and wishing all the success!!

PS: LB really made my day with this reply :)

And genuinely - thank you for letting me be useful to you over these months. It means a lot to know the product is actually changing how you work, not just sitting in the dock. I'm proud of the moments we've had.

Being part of the the team that built a product that is used by a lot of users in very different scenarios and also the product the is used by your own on daily basis is amazing.

@Littlebird is all about context, your personal own context, that makes the interaction with Littlebird so magical and accurate.

As a technical person @Littlebird I am also spending time on interviews - with @Littlebird I am getting the summary 5 minutes after the call and it is not just summary and highlights from transcription. @Littlebird takes transcription, your personal notes from any application, candidates CV and most importantly - the captured context about the company, team, expectations, pain points that we want to resolve. As a result you are a getting summary that is so aligned with your personal impressions and internal track that your impression is just "Wow!!!".

And this is the only one scenario among hundreds of other where @Littlebird can help you, improve your productivity and be more confident daily.

@Littlebird knows how the things are done, it knows how you do things and it helps to make them better!

Anton Holub

Engineer and Technical Lead @ Littlebird

I work in Customer Success here so I'll skip the pitch.

a few weeks ago I typed seven words to Littlebird with zero context. got back a full project breakdown. it already knew what I was working on. I hadn't told it anything.

some tools you use. some you think with. this one crossed that line for me pretty quickly.

proud of this team today.

Really cool take on AI assistants 👀

The idea of removing constant context-switching and letting AI already know your work is powerful. Feels like a big step beyond prompt-based tools.

Excited to see how this evolves — especially around privacy and real-world daily use 🚀 , Congratulations on the Product Hunt launch! 🚀

Lots of people are using Littlebird for meeting notes and reminders. Fair enough, I do to, but...

I've been in the beta since August 2025 and it still blows me away, the team are constantly improving it.
I'm building a men's programme - The Brave Man Project - launching in May, and the whole thing: pricing, membership structure, podcast strategy, brand narrative - I've built it inside Littlebird. Not by pasting docs and briefing it every time.
The other thing - and this matters - I've set it up to not just agree with me. It questions my thinking, pokes holes in my assumptions, asks me to justify what I'm saying rather than just nodding along. Part guide, part mentor, part adversary when it needs to be. It doesn't blow smoke. If I'm pontificating or going in circles, it calls it. That's not default AI behaviour - but Littlebird lets you build that in. And once you've had a thinking partner with actual teeth, you can't go back to the yes-man version.
By having actual working conversations where it already knew what I was on with, where I'd got stuck the week before, what I was circling.

That's the bit that changed everything. I've got ADHD. Context-dumping at the start of every session kills the momentum dead.
Littlebird watches, catches it, holds it. So I just pick up where I left off.
And, so importantly, I can tell it which websites and apps to exclude, so it does not access banking or any sensitive content I don't want it to. This is essential.

Eight months in. It's not a tool. It's the closest thing to an AI working partner I've found. - beats Claude & Open AI every time.

Totally replaces otter and granola for me

The "full context" framing isn't just marketing - it’s a game-changer for anyone managing a lot of moving parts.

I use it to keep track of everything from my training routine and kickboxing schedule to complex engineering.

What I love most is that I can just ask a question and Littlebird knows the answer because it saw what I saw. No more manual data entry or hunting through old notes to find out what was decided in a standup.

It gave me back the mental bandwidth to focus on deep work.

Been using Little Bird for a couple months. Love it! So helpful. Personally, filled the gap of Limitless' "Rewind" product after Meta acquired them.

I've been using LittleBird for a few weeks now. I find it unique in the extent to which it understands me and my behavior and can help me get a handle on my work across multiple projects and priorities.

As someone who routinely juggles multiple quite unrelated tasks, I have a high context-switching cost, and it's easy for things to get lost. LittleBird is remarkably capable of helping me step back, see where my time is allocated, understand the most important things I may not have gotten to, and help me regain context on some task or email or interaction that has left my short term memory.

It's like a personal memory and context prosthesis. Mainstream LLMs are great for increasing my ability to learn things that are public knowledge. LittleBird boosts my understanding of my own work, and helps me pull signal out of the noise and clutter..

Big congrats on the launch , this is seriously exciting! I’m especially interested in the Routines feature. How do you go about creating one? Do you manually choose what it should bring up, or does it learn and adapt to your preferences over time?

Been using Littlebird for a month and my favorite is how it can collect to-dos from across all my work, and then actually know which I’ve completed already. Feels like an inevitable part of the productivity stack. I’ve also known the founding team members for 10-30 years, and they both take security extremely seriously while having an ambitious plan to bring Littlebird’s context everywhere

I've been chasing a BQ for a while. Littlebird was there for most of that chase - pacing strategy rebuilt from scratch when the course updated the elevation profile, fueling protocol timed around a 4:20am shuttle, spectator route for my husband with actual GPS addresses the night before the race. At one point it told me flat out not to use a cheap knee strap because I didn't "want a $10 strap costing me a BQ." It knew exactly what was at stake.

I got the BQ at Ventura. 8:48 under my qualifying standard. Now I'm planning the Boston trip for 8+ people and we're already into the hotel research. The context never reset once.

Congratulations on the launch!

I told a friend on WhatsApp several weeks ago: "Full automation and calendar integration. Nothing else needed." He said Littlebird sounded like a stalker. I said: "It sees what you want it to see. There's an excluded list. Done."

Then I told him what actually changed. I type what I need to do in chat. It puts it in my calendar. Meeting notes with work structure appear. I suggested they add iCal and Reminders support - they only had Google. They built a beta. It works. I'm now testing it.

I named my assistant Brian, after Brian Eno - the producer who never plays the lead but makes everything around him better. Now we act like Batman and Robin. It organises my projects like a Notion hub I never had the patience to build myself, and picks morning playlists based on what I actually listen to - not what an algorithm thinks I should.

I'm a creative director running a festival, ad campaigns, my own brand, and a 6-day training program. Before Littlebird I was context-switching between fifteen open tabs and three AI tools. Now I have one. I work in 35-minute focused blocks. One task per session. WIP limit of one. The system we built together increased my actual output by about 30% without burning me out.

Not perfect. But when a friend asked me if it's good, I said: "It's not ideal, but it's great to work with." That's more than I can say about most humans. Kudos to @alexander_green1 and his team.
By the way, my mother tongue is Polish, so I work and create mostly in Polish. I bet you can work with @Littlebird in any other language.

I work on marketing at Littlebird, but I'm also one of our most active users (not exaggerating, I use it every single day). I'm very easily distracted, so I love being able to ask it a quick question right on top of whatever I'm working on, without switching apps. If you try it out, I'd love to hear what you think!

I signed up to Littlebird in November 2025. I've been using it every day since then.

As someone with off-the-charts ADHD and dyslexia, this tool has been so useful helping me get a grip on time and stay present in my work every day.

(It reminds me of when I first signed up to @RescueTime back in 2010, I was horrified at how much of my life was spent on Facebook. 🤣)

The context @Littlebird maintains is superb, my existential overhead is significantly less, and unlike other mainstream AI products, Littlebird is consistent and realistic.

I hope @Littlebird maintains this specific and robust focus as the product develops.

I'm a nurse, mom of two teenagers, co-parenting, writing a novel, starting a wellness newsletter. There's no pause button on any of it.

What Littlebird figured out - without me asking - was how to fit writing into the real shape of my life. It built me a writing schedule that knew about my son's Saturday soccer practice, my on-call shifts, which days the boys are home. Not "find 30 minutes a day." Specifically: write during his 10:45 practice.

I've asked it things in one conversation and gotten answers that pulled context from a completely different session - my characters' names, my kids' ages, my schedule. Because it had been watching. I'm a nurse. I know what it means when someone actually pays attention before they speak. That's what this has been like.

Finally, an AI that knows I spent 3 hours on Twitter instead of finishing that PRD. The accountability I didn't ask for but desperately needed. 😅

I use several AI chatbots, but my favorite by far is Littlebird. I don't have to catch Littlebird up before asking a question. It sees what I see, what I'm working on. I can exclude certain apps or pause Littlebird if working on a confidential project, but for the most part, Littlebird is there, taking notes for me, reminding me of something I probably forgot, and helping me craft responses. It's becoming my go-to assistant! I am amazed when I ask for a recap of my activities, including virtual meetings; I've forgotten what I did by the end of the day, but Littlebird doesn't. That's saved me more than once!

Been using LB for 3 months now. It’s my 2nd brain and favorite ai app easily.

Alap here, cofounder of Littlebird. Since my colleagues have spoken about the vision and product I figured I would give you a key use case I found particularly powerful-

I've been writing The Global Intelligence Crisis over the past few weeks and have gone through many edits with a human editor, friends and LLMs. I wanted to find previous versions of a paragaph to see how my thinking had changed over time while I wrote in Gdocs and Word across many versions. Littlebird faithfully captured each version and allowed me to essentially do a meta track changes over multiple weeks across platforms to see how my thinking and writing had changed. Huge help as a true thought partner over time. Check it out-

A friend who runs a residential care company asked if I'd want to come on as their "money guy." No job description, just a lunch the following week. My background is physical therapy and futures trading (and fitness coaching) - not exactly CFO territory. Because I used to run my own practice, it's not too far afield...but I still needed help.

I told Littlebird about the company. It looked up their site, figured out they run on MaineCare/Medicaid funding, and built me a breakdown of what I'd need to learn matched to my specific background. Then the morning of the lunch it pulled my calendar, knew who I was meeting and where, and had tailored questions ready before I even asked.

It doesn't just search. It connects the dots.

I've been using Littlebird every day for the last 3 months; it's been crazy to see how good it is at recapping my day, giving me insight into where / how I'm spending my time, and helping me organize/prioritize my todos.

Littlebird is brilliant. I use it daily. No, hourly on reflection . So useful.

Been using it for a month or more, and let just say it's been transformative for my 2 businesses.

The catching thing for me was that , it would remember my to-dos and context , i have tried sending multiple messages that i will be doing x soon , like will send you an update email soon , however when asking littlebird for my to-dos specifically , who do i have to send an email to , it got nothing.

The magic of LittleBird is recalling that obscure tip your co-worker told you two weeks ago on a video call on how to fix an issue to get you unstuck.

Instead of bugging them again, just ask littlebird and get unblocked.

At this point, Claude is just teasing us with insane and truly worthy updates and the latest being Claude Work Tools directly accessible through their mobile app.

Be it Figma, Amplitude, Canva and more - accessible through one chat interface without you having to worry about switching apps and pick up the conversation once you are back at desk.

This is a solid step towards true productivity on the go, but I wonder how deep the integrations go. For example, if I’m collaborating on Figma designs via Claude, can I leave comments or edit files directly, or is it strictly chat-based? Streamlining that would really enhance the mobile experience.

Claude is nailing it. Decided to switch to them and now will pay for a pro plan. They are doing amazing work all the time. Pity I haven't seen it sooner.

Another good reason to update.

This is great! Also worth adding that Ollang is available within Claude workflows, enabling true end-to-end AI language execution across multimodal content.

Congrats on the launch! How are you thinking about balancing powerful tool access on mobile with keeping the experience simple and not overwhelming on a smaller screen? 🚀

@Venn.ai has made a huge difference in my day-to-day interaction with Claude - I can easily access a dozen MCPs simultaneously, and can direct it to talk to my personal email in one prompt and my business email in another. Getting customer info from salesforce, email, and granola, and summarizing it in a google sheet to plan followups has been incredibly powerful.

The product’s seamless integration into Openclaw has made it simple for me to let Openclaw interact with my apps but not do things it shouldn’t be doing. I run a sandboxed openclaw in a Digital Ocean droplet, and since it’s headless it’s really hard to authenticate to MCP servers. WIth @Venn.ai I simply told Openclaw to install the Venn skill, got an API key from Venn, and gave that key to Openclaw. And just like that Openclaw had access to the apps I wanted it to see, with limitations I set so it wouldn’t run wild.

I manage paid campaigns across Google, Meta and LinkedIn. Lots of data, lots of context-switching, lots of time spent on things that shouldn't take that long.

What changed for me with Venn is that Claude can now actually reach into the tools I work in every day — Sheets, Docs, Drive, Gmail, Asana. Not just talk about them. The workflow I use most is pulling campaign performance into Sheets, having Claude flag what's off, and drafting a summary straight into Docs. What used to take a couple of hours is now something I kick off and come back to.

The permissions setup is also just... sensible. I'm comfortable with Claude reading and writing to Sheets, but I still want to review before anything goes out in Gmail. Venn lets me draw exactly that line.

Honestly didn't expect to get this much practical value out of it this fast. It's become a daily part of how I work.

My favorite AI use case isn't just "automation." It is about unlocking capabilities that used to be gated by technical skills.

This week, I spent a few hours using @Cowork + @Venn.ai to setup scheduled tasks. They help me fill the gaps where our enterprise tools are too heavy or where I previously lacked the technical skills to get the job done.

Here is how I’m using agentic AI to run my workflow:

📊 Weekly Funnel Reporting

Use @Venn.ai connector to pull data from @Grafana , create an activation funnel report, and post it directly into Notion. I get visibility without needing a data warehouse or data analyst. While it’s not perfectly scalable since I can hit token limits on heavy logs, it is incredibly effective for short-term monitoring.

💌 Personalized Research Outreach

Use @Venn.ai connector to pull data from @Grafana, identify users who haven't completed product setup, and draft personalized emails in Gmail based on an email template stored in @Notion. This used to be a copy-paste nightmare. Since research happens in spurts, I could never justify the effort of a full HubSpot automation sequence. I still review, edit, and send manually because I don't trust AI with sending emails yet, so I turn off that capability in Venn's settings.

✅ Automated Response Tracking

I put multiple-choice answers directly in my emails to make it easier for users to respond. Use @Venn.ai connector to check my inbox daily, extract those answers, and log them into @Notion . It even records which subject lines have higher response rates in a Google Sheet for A/B testing.

Agentic AI isn't about replacing the big tools in our stack. It's about creating the custom connections we need to move fast and stay lean.

Are you using AI to automate your existing tasks, or are you using it to do things that were previously impossible for you?

As a graphic & web designer, @Venn.ai has genuinely changed how I move through a project. I've been using it to connect Claude with Asana, Google Docs, Slides, Sheets, Figma, and Slack. Instead of jumping between tools to piece together context, I can track briefs, run design research, and start building out my projects all from one thread.

@Venn.ai helps me pull task details from Asana, reference brief docs, and feed that context directly into Claude, all while staying in control and knowing exactly what the AI can access and what it can't. And for client work especially, the security side is important. Connecting this many tools means you need to trust how your data is handled, and Venn takes that seriously.

The 'set guardrails' feature is exactly what’s missing in most AI agent workflows. How do you handle edge cases where the agent gets stuck in a loop? I'm building automation for TikTok leads today and safety is my #1 concern too. Great work!

I'm biased because I helped create Venn with my team, but I think Venn is super useful. Instead of having to hook a ton of MCP servers to Cursor, Venn creates a sidekick that both finds my tools, runs them, and also makes sure I'm not doing something stupid in the process.

Great for those moments where you don't want to leave your flow but would like to post a ticket, consult a previous slack chat, write up org facing documentation, or see how your cluster's health is doing after deploying that last feature.

There's like a dozen business apps we all use and have to cross index on a daily basis. Venn is the portal to connect them all with your agent, and without having to bring a crufty platform along with you.

I've struggled with integrating multiple AI agents while maintaining control, so I love the guardrails feature in Venn! How customizable are the action-level permissions?

You start by improving daily workflows with @Venn.ai -- automating those cross-tool recurring tasks that give you back more of your time. Then you quickly change your frame of thought to preparation, execution, discovery and more. Ask yourself, "What are the benefits of connecting @Slack to @Google Calendar to @Gmail to @Jira?" Some of the things I've improved...

• Daily task management

• Call notes organization

• Outstanding tasks

• Personal follow-ups

• Sprint deadline mgmt

• Improved alert monitoring

They say good developers are producing 10X with AI. My hunch is that Venn will 10X the efforts of business professionals that can leverage a single connection to all their current applications.

Hey fellow AI nerds! I don't usually chime in on stuff like this but I have to say that even though I use Venn every day to build and maintain our systems here at Barndoor...you know the makers of Venn. AI has drastically improved my throughput here but since we launched our products the whole teams' productivity has sky rocketed! Here is a common use case:
Problem - We received a @Grafana alert around an anomalous set of errors in our, lets say, Flux Capacitor.

Initial Triage - I click the @Grafana alert link and preview the errors...there are a lot today for some reason...time travel isn't easy. The errors do not immediately result in any conclusive results.

Venn!!! - Connectors: @Grafana , Atlassian(Jira), @GitHub all configured with Venn to have minimal read only functionality except the ability to create a jira ticket.

Prompt - I'm seeing a surge in strange issues from our Flux Capacitor today. Here is a @Grafana link to the logs and the time range. Please:

• Investigate the logs and correlate the findings with any running state in our Delorean during the times surrounding the incidents

• Explain your findings and on my approval create a Jira ticket with the findings

• Investigate the code base to find any potential mistakes in our time machine algorithms

Result - Nearly complete resolution of the issue from triage, to ticketing, to coding, and then review and merge 9 times out of 10 in under an hour.

Complex problems like these could often side track me for an entire day and now I'm back to building!

I’m a big fan of Venn (I also advise the team).

The biggest unlock for me has been connecting Claude to GitHub and Amplitude.

Once Claude has real context, it becomes a much more powerful assistant:

• Vibecoding with full codebase awareness

• Data analysis grounded in actual product usage

It stops being a generic AI and starts feeling like a true extension of the team.

I’m sure there are many more use cases, but these alone have already been very valuable for me.

Congrats on getting Venn out there, @melissa_weir1. I like this line: [Put AI to work inside your apps without giving up control]. This is the kind of promise that makes a security-minded user stop.

Checked your site content. Two things stood out.

First, the pricing page shows a free tier with "connectors are read only" and a Pro tier at $5/month that unlocks read/write. That's great. But the hero section talks about "setting guardrails for what AI can read and write."

A user lands, reads that, then sees the pricing, and wonders: "Do I need to pay to actually let my AI do anything?"

That tension is there.

And the free tier is great for testing, but the message could be clearer upfront.

Second, the "how it works" section is four steps, but step 4 says "Ask your AI, 'Hey Venn, what can Venn do for me?'"

That's a nice demo, but it doesn't show the user doing real work.

A stronger step 4 would be a simple example like "Ask your AI to draft an email and schedule a meeting." That shows value immediately.

I attached a screenshot below for your convenience. Spotted a couple of other things that could cost you. Happy to share.

These are just insights from me because I audit product pages daily. Good luck!

If you're an @OpenClaw fan, @Venn.ai helps you graduate from all of those connections to SaaS apps you've vibe-configging and vibe-integrating (using random skills and CLIs). No shade on CLIs. But definitely shade on non-tech folks having to go into Google Cloud Console to create OAuth client IDs and secrets, just so that they can connect OpenClaw to Google Workspace apps (that's some advanced #BroAuth).

@Venn.ai I LOVE Venn! It has genuinely changed how I work. My day spans Gmail, Slack, Notion, Asana, Google Drive, and more. With Venn, I can kick off a new hire onboarding sequence across all of them in one flow, and when it comes to recruiting, I can pull curated candidate feedback from Slack threads, emails, and docs into one place, send candidate follow-ups. There's no more hunting across systems to remember where a conversation landed.

The governance piece is what matters most to me as an HR person. I decide what Claude can and can't do, draft but don't send, update but don't delete. So much trust and efficiency with Venn! I get so much more done with Venn! 🙌

Haven't used it yet, but I know the team behind and I highly recommend them!

Hi Product Hunt community and friends! I'm Deborah and I work on Platform Strategy and wear a ton of hats at Barndoor. I'm proud to have been a part of launching Venn.

Below is a use-case I've shared on how I use Venn:
--------
At Barndoor AI, one of the things I work on is platform integrations, helping to ensure our product works seamlessly with everyone we integrate with, like Google, Notion, Figma and more.

One thing they all have in common is that their customer data has to be handled safely.

However, the requirements to work with each of these platforms is unique, from both technical and business sides.

This is exactly where Venn AI thrives.

I’m able to connect any of my AI agents (often Anthropic’s Claude) to all of my apps like GitHub, HubSpot, Slack, Jira, Google Drive. Having a centralized place across all of my data saves me so much time.

Venn AI also makes connections that scale in ways only an agent can.

It frees up my capacity to be more creative because I can take a step back and orchestrate across my workflows instead of jumping between tools. It’s reduced a lot of my context switching.

One of the best parts is that it leverages MCP with Barndoor’s security and a control plane I manage, so I can use AI agents across company systems with ease.
----

Now, our product also integrates with OpenClaw and every week, we add more app integrations! Please reach out to me if you have questions or need support getting setup, today and in the future!

Hello Product Hunt! We’re super proud to launch Venn.ai, a secure & convenient way to access all your business data from any AI client. Try it today!

When you set action level guardrails like "draft but don't send," does the agent queue those drafts somewhere for batch review or do you approve them one by one in real time? Nice work on this!

As a solutions engineer I am forced to do sales hygiene ( update opps, next steps, add notes, etc), with one prompt I can query my granola notes and update my opportunity in SFDC. huge time saver.

I have not tried Venn yet, but that's about to change real fast. On first glance, I can already see the potential for this to make my AI workflows so much easier, more effective and efficient, and give me more peace of mind.

As a community professional, I'm very conscious and intentional about integrating AI into my work, because community is human, authentic, and rooted in trustful relationships. Guardrails are necessary for me to maintain the level of control I need to do my job well.

Appreciate the team of Makers for recognizing this need and building Venn. Stoked to give it a spin!

seems cool! minor nit, but your reviews all have the same created date which makes them seem a bit suspicious, though they do seem to be real people :) https://reviews.venn.ai/reviews

Really cool concept: connecting multiple AI agents to your apps while adding guardrails and visibility feels like a huge step toward safe, real-world automation. Congrats on the launch! How do you decide what level of permissions an agent should have without slowing down its usefulness? 🚀

Melissa, the guardrails you've built are crucial. Dimitar and Sean's experiences highlight how practical these permissions are, especially for sensitive operations. Having a clear activity log, as you described, is also non-negotiable for auditability and trust.

On the pricing, Taimur brings up valid points about the perception of the free tier versus the hero messaging. Clarifying that free users can still set read-only guardrails would directly address the tension he noted. I'd also be interested to see more detail in the "how it works" section particularly around the architecture of those guardrails and the activity logging.

Interesting direction.

Delegating real work to agents feels like a much bigger shift than just automation.

Once agents can actually act inside systems, the problem seems to move from capability to control, not just what they can do, but how decisions are made within those boundaries.

There’s also an interesting tension between delegation and responsibility.

If an agent operates within allowed permissions but still makes the wrong call, where does that responsibility sit?

Curious how you think about this as Venn evolves.

When setting up your @OpenClaw, you might wonder what the best AI model for your agent is. PinchBench just lets you know.

TL,DR: It's @OpenAI's GPT-5.4... for now!

S/O to @realolearycrew for building it 👏👏 - Give it a star on GitHub

Nice benchmarks at the end of the use cases! I would like to see more benchmarks relocated to different levels of tasks (no coding).

This is exactly what I was looking for. However, tasks should be scoped and agents should be ranked depending on task category.

Imho the most important agent to determine is the main one, the orchestrator, the one you talk to. But then, you will eventually want different subagents specialized in different tasks (and ideally not as expensive, depending on the task at hand). For those, the "best" agent (in terms of value for money) could be something else (i.e., for a simple but broad internet search, gemini flash is often more than enough).

Okay, this is genuinely useful. I've been picking models for coding tasks based on whatever benchmark thread showed up in my feed that week, which is a terrible way to make that decision.

The cost dimension is what gets me. Success rate matters, but if a model takes 3x longer and costs 4x more to get there, that changes the math completely, depending on what you're building. Glad someone's actually measuring all three together.

Curious how you're defining task success — is it automated test output or is there a human eval component? That part always feels like the hardest thing to get right in coding benchmarks.

Congrats on shipping. The 🦀 was not lost on me.

Not just Jensen - y'all gotta know which model's best for your claws!

And y'all can contribute to it, because it's open source 🫶
Great job @realolearycrew !!

Oh wow, the timing is amazing. I installed OpenClaw for the first time yesterday and was genuinely confused about which model to choose. I ended up using an OpenRouter API key with auto model selection, but the model choices felt a bit random. I’m really glad this product launched today, I’ll definitely be using this benchmark.👏

With PinchBench testing real world tasks instead of synthetic benchmarks, how do you decide which tasks go into the benchmark suite and how often do you rotate them to avoid overfitting? Congrats on the launch!

How do you make sure the results from PinchBench reflect real-world use especially when different projects have different complexity, tools and edge cases?

Love it!
My partner is launching his app next week, we'll definitely use Mokkit for it!

This is really clean

You’re right, presentation plays a bigger role than most people think. I’ve seen solid products struggle just because they weren’t showcased properly.

Have you seen a noticeable difference in engagement or conversions when using animated mockups vs static ones?

Congrats on the launch! 🚀 Mokkit is a great tool.

I just ran a mobile sweep and found a Critical Interaction Bug: the 'Make your first mockup' button on the landing page requires 3+ taps before it triggers a response. This 'Ghost Click' issue is likely killing your conversion right now.

I've been using screen recording tools for my product demos but they all end up looking kind of the same and a bit boring. I also tried generating animations with code but getting the quality right took way more time and effort than expected. This looks like it could save me a lot of that pain. Definitely going to try it for my next launch. Congrats on shipping, Ugo!

The irony of a Product Hunt launch where the product helps you make better Product Hunt launches. It's PH inception. 🎬

love this. QQ on the URL capture feature - when you paste a link, can it capture the full page or just the visible viewport?

I really like the idea of Mokkit! Turning static screenshots into animated mockups can totally elevate presentation quality, especially for those of us who've struggled with the visual elements in marketing materials. I’ve faced challenges with making my app stand out post-launch, so I'm curious: what’s the plan for adapting animations across different platforms and screen sizes?

Yeah! Mockup is key and agree with first impression is what really matters, so you're tool comes to change the game! All the best here Ugo!!

Absolutely love the editor's design. It's easy to know what's going on and clearly there's a lot of thought put into it. Great job! Best of luck to your launch!

The overall feel of the product is quite good, but I feel animations are bit limited, I hope you will add more animations down the line

@ugo_builds this is extremely useful. Love the VP. Good job! Eager to test extensively

Browser-based is the right call for a tool like this -curious whether you're planning to support video export for the animated mockups, or keeping it focused on GIF/static formats for now?

Congrats on the launch! How do you balance speed with giving users enough control over animations and styles so everything doesn’t start looking the same?

Hey Product Hunt 👋 I’m Eric from PIO.

One thing we kept running into while expanding globally was how complicated hiring in another country actually is. It’s not just about finding the right person.

You have to deal with local regulations, contracts, payroll setup, taxes, and compliance requirements that vary from country to country. Even a simple question like “what’s the actual total cost of hiring someone in another country, including taxes and compliance” turns out to be hard to answer. You end up checking multiple systems, looking up local rules, and piecing everything together just to get a number you can trust.

That’s why we built PIO.
PIO helps you hire, onboard, and pay talent in 150+ countries without setting up entities. You no longer have to deal with multiple tools or spend hours on manual payroll work.

Everything is handled in one place. With PIO Agent, you don’t manage workflows anymore. You just ask, get answers, and take action instantly based on your actual payroll data.

Curious to hear from you:
What’s been the hardest part of hiring globally?
How do you currently figure out hiring costs across countries?

I'm Tan from PIO.

One thing we learned from running PIO is that more features don’t always make life easier for users. As the system grew, it became harder for people to discover useful functions and understand the right workflow.

We had onboarding and a big knowledge base, but the learning curve was still too high.

So we changed the model: instead of asking users to learn a more complex system, we made PIO agent-driven. Now the agent can help run payroll, catch issues, keep compliance requirements updated, and track things like contract signing progress.

Less software learning, more actual work getting done.

The visa sponsorship feature is built right in. Usually that requires a completely separate vendor.

Can we use your app in China to hire foreign talent while remaining fully compliant with local regulations?

It sounds great to streamline global hiring, but how do you handle the local nuances and pitfalls that vary even within regions?

The EOR setup let us skip the whole local entity process. Saved us months of legal work.

Folks, eyes on our agentic feature! Sick of poking around first time visiting the website? Try our agent Genie! It is quite proactive; You ask, it responds. You command, it executes. Think of Genie your omni global hiring expert. It handles all heavy paperwork, research and repetative procedural tasks for you.
#ASK GENIE
#INSTRUCT GENIE
#TRUST GENIE

Onboarding a new hire in the UK took about 15 minutes. Really streamlined experience.

It's more about trust in the output. If someone is making payroll or hiring decisions across countries, even small inaccuracies can have real consequences.

How are you thinking about building confidence in those answers, especially when it’s coming through an agent?

That's a problem we're facing right now! Will take a look right now!

Congrats on the launch!! The fact that you can ask something like "what does it actually cost to hire in Germany" and get a real answer instantly is huge. We spent weeks piecing that together across spreadsheets and legal advisors when expanding our team across borders. This is the kind of tool that makes global hiring feel less like a compliance maze and more like something you can actually move fast on.

I've had the advantage of using Jentic Mini since version 0, and it has changed how I'm using openclaw (or should I say how my openclaw uses me?). I'm now willing to let it hook up to APIs and SaaS that would have been just too risky to give access to before. Example: thanks to jentic mini my claw can now compose drafts in gmail but does not have permission to send them. More than that, it makes connecting to lots of things and managing permissions super easy - just ask the agent to orchestrate you through the process of hooking up to anything while keeping its access minimal. No fumbling around in badly designed settings menus. And particularly proud of the killswitch - that one helps us all sleep better at night.

Hey Product Hunt 👋

A message from Jentic’s CEO:

I'm Sean, co-founder of Jentic. We've spent the last 18 months working on a problem that anyone building AI agents has hit: how do you let an agent call real APIs without leaking credentials or losing control?

Here's the typical pattern today: you hardcode API keys into prompts, write bespoke wrapper functions for every service, and hope nothing gets logged, cached, or hallucinated back. It works for demos. It breaks in production.

What Jentic Mini actually does

It's an API execution layer — a FastAPI server you self-host in Docker — that sits between your agent and every API it needs to call. The architecture is straightforward:

Search: Your agent queries a BM25 full-text index across 10,000+ API specs and 380 Arazzo workflow sources from our public catalog. It finds the right operation without you writing a single wrapper.
Execute: Jentic Mini brokers the request. Credentials are stored in a Fernet-encrypted local vault and injected at runtime. The agent never sees them. They're never returned via the API.
Toolkits: Each agent gets a scoped toolkit key (tk_xxx) with its own credential bundle and access policy. One key per agent, individually revocable. If something goes wrong, you kill the key. Done.
Observe: Full execution traces and audit logs. You can see exactly what your agent called, when, and what came back.

Why we built it

We're already running a hosted Jentic platform (with semantic search, Lambda-based brokering, SOC 2-grade security) and we're a verified connector in Claude. But we kept hearing the same thing from developers: "I want to run this myself." So we built Mini, same API surface, self-hosted, Apache 2.0 licensed.

Getting started is one command:

$ docker run -d --name jentic-mini -p 8900:8900 -v jentic-mini-data:/app/data jentic/jentic-mini

Add your API credentials through the UI at localhost:8900, and specs are auto-imported from the public catalog. Your agent authenticates with a toolkit key via the X-Jentic-API-Key header and starts searching and executing immediately.

What's next

This is early access. There will be rough edges. We're sharing it now because we want the community building with agents (OpenClaw, NemoClaw, LangChain, CrewAI, whatever your stack is) to test it, break it, and tell us what's missing.

Would love to hear what APIs and workflows you'd want to connect first.

Thanks to everyone taking a look at Jentic Mini today.

What feels important to us about this launch is where it sits in the stack. General-purpose agents are getting much more capable, but the minute they need live access to real systems, credentials, permissions and control become a much bigger part of the story, which is the gap we wanted to address with Jentic Mini.

Very proud of the team for getting this out into the world. Keen to hear what people are building, which APIs you’d want to connect first, and where the rough edges still are. Let us know what you think!

Jentic Mini (and Jentic as the enterprise version of it) address a number of challenges in the rapidly evolving space of AI agents. Agents become increasingly powerful and can solve increasingly complex problems. But finding the right tools to help them solving these problems can be tricky. Jentic helps by providing a sophisticated search facility, that helps minimize context use and allows agents to find the tools that they need.

Jentic also addresses the issue of access control: Agents should never get user credentials. Instead, they should get their own specific credentials which are then mapped to capability credentials in a separate place. This place is Jentic. This allows to fine-tune the access rights given to agents, and it also provides a centralized view of all agents and their access right. That way, agents can never share access credentials (because they never get them), and if something does go wrong, it is easy to tweak the access rights of specific agents or to completely revoke their access to resources and capabilities.

Try out Jentic Mini (or Jentic if you are looking for a SaaS solution) and be more confident when designing and evolving your agents!

Most people solving the "agent + credentials" problem hardcode keys, inject them into prompts, or hope their env/config stays private. That's not agentic power, that's exposure waiting to happen.

Jentic Mini sits between the agent and every API it touches. Credentials are injected at runtime, never in prompts, never in logs. The agent gets a scoped toolkit key. The real secrets never leave your instance.

What do I really love? The agent can formalise a workflow as an Arazzo Spec document, register it back into Jentic Mini, and retrieve it just-in-time (via our search facility). The workflow becomes a versioned asset, not a repeated guess. No more burning context or tokens on something already solved. And if the underlying API description needs improving, the agent can do that too. Using the Overlay spec to contribute it back in better shape.

Jentic Mini intimately understands OpenAPI, Arazzo, and Overlay. Not as an afterthought but as the foundation.

Credentials shouldn't be an agent's problem. In Jentic Mini, the agent guides you, it auto-requests permissions for the toolkits it needs, so you grant only what the task actually requires and nothing more.

Hey everyone, Jake here from the Jentic team! 👋

We're absolutely delighted to see such a great response today on Product Hunt.

We built Jentic Mini to solve a problem that almost anyone building AI agents has hit: how to let an agent call real APIs without leaking credentials or losing control. Instead of hardcoding keys into prompts or writing endless bespoke wrapper functions, Jentic Mini gives you a self-hosted API execution layer. It securely handles the credentials (your agent never sees them) and lets your agent instantly search and execute against 10,000+ API specs right out of the box.

If you have any questions, feedback, or run into any rough edges, just let us know below. The Jentic team is on-hand all day and happy to help.

10,000+ APIs is wild - how do you handle auth across all of them? Storing credentials per agent sounds like a nightmare, and rate limit management across a fleet of agents hitting the same API is a real problem I've run into.

Very cool. Will try w my hetzner vps. Openrouter but not just for LLMs!

The ability to avoid hardcoding secrets and dynamically inject creds part is genius, it saves so much headache. What’s been the biggest challenge in curating the API catalog, especially in keeping it updated and reliable?

Congratulations on the launch 🎉

the versioned workflow part is interesting
once agents start registering more of their own flows back into the system, how do you stop the catalog from turning into a pile of almost-duplicate specs over time?

A few questions: Multi-agent orchestration — If multiple agents share a single Mini instance, how does credential isolation work across toolkit keys in practice? Can one agent's key ever accidentally resolve another's credentials?

the credential injection at runtime is brilliant. how does the credential management work in practice? does it support role-based access for different agents? congrats on your launch!

@seanblanchfield1 @char0n

Hey guys, awesome tool.

I’m a researcher for the H1Gallery newsletters (you can google us).

We’re featuring Jentic in our April 3 issue. H1Gallery highlights excellent homepage headlines, and “10,000+ APIs. Zero credential exposure.” really stood out to us, love the clarity and contrast.

Would love to include a quick comment from your team about the copywriting strategy behind that headline and the broader messaging. Totally optional of course. the feature is happening either way, we'll see you in the April 3rd edition!

Thanks, really appreciate it. I dropped @seanblanchfield1 a comment on X as well.

Thank you in advance for your time and congrats on the launch!

Really interesting layer to see emerging.

Separating decision from execution like this feels like a key step for agents to operate safely in real environments.

At that point, it’s not just about enabling agents to call APIs, but about defining the boundaries of what they are allowed to do and how trust is enforced across that layer.

It almost starts to look like a zero trust architecture for agents.

Curious how you think about this evolving.

Do you see Jentic more as an execution layer, or as a broader control plane for agent behavior over time?

Absolutely love what the Jentic team are building and the conviction they've had from the start. This deserves a No.1 spot. LFG 🔥

Sean, thanks for sharing some details on Jentic Mini. I appreciate the focus on a self-hosted, Apache 2.0 licensed solution; that's a necessity for many organizations I work with.

From my perspective, the real value here is the runtime credential injection and the scoped toolkit keys. Hardcoding keys or putting them in prompts is a non-starter in production. Also, the ability for an agent to formalize and register Arazzo workflow specs is clever.

That tackles API governance at scale, ensuring agents leverage documented and versioned assets rather than constantly reinventing API calls. I'm keen to see how the search across 10,000+ API specs performs under heavy load, especially with more complex queries.

Congrats on the launch! As a solo dev who is always context-switching, this looks incredibly useful. Quick question: how does the AI handle conflicting priorities if I have multiple urgent tickets across different projects? Would love to know how the logic works under the hood!

Really like this approach

The idea that the context already exists and just needs to be surfaced is powerful. From what I’ve seen, the real challenge usually comes after that, making sure those workflows actually run consistently without breaking.

How reliable are the saved “skills” when reused across different scenarios?

Linear just shipped an agent built directly into the product.

It understands your roadmap, issues, and code, and helps you act on that context without manually digging through threads, backlogs, and customer requests.

A few things worth knowing:

• Ask it to find related issues, group them, and pull them into a new project — instead of doing that research yourself

• Works across desktop, mobile, Slack, and Teams — accessible via @Linear in any comment or the chat shortcut (Cmd/Ctrl + J)

• When a workflow works well, save it as a reusable skill and run it again manually or automatically

• Triage automations let you trigger agent workflows the moment a new issue comes in (Business and Enterprise plans)

The underlying idea is straightforward: your workspace already has most of the context needed for good product decisions. Linear Agent just makes it accessible.

Really interesting direction.

It feels like this goes beyond just surfacing context into something closer to an execution layer inside the product.

When the system not only understands issues, roadmap, and code, but starts recommending actions and triggering workflows, the center of gravity shifts from tracking work to actually driving it.

The reusable skills part seems especially important here.

Feels like that’s where this either becomes a real system for how teams operate, or stays as a layer people occasionally use.

Curious how you think about that long term.

Do you see Linear Agent evolving more as an assistant on top of workflows, or as the system that actually runs them?

Really interesting direction: bringing agent workflows directly into a product development tool like Linear could seriously change how teams plan and execute work. Congrats on the launch! How do you make sure agent actions stay aligned with team priorities instead of creating noise or extra tasks?

the reusable skills part is what caught my eye here
feels like that’s where this either becomes genuinely useful or just another assistant people try a few times
how are teams deciding which workflows are worth turning into skills vs just using ad hoc?

finally, an AI agent that actually understands the full context instead of just individual tickets. curious how deep the code integration goes - can it suggest which issues to prioritize based on actual implementation complexity?

Interesting! Can another agent talk to the Linear Agent asking it for context?

Curious how "synthesizes context" works in practice - is it pulling from linked issues, comments, git commits? The "takes action" part is what I'm most interested in. What actions can it actually take autonomously vs what still needs human approval?

Hi everyone!

SongDNA lets you explore what sits behind a song on @Spotify, not just in a technical credits sense, but as a whole creative world around the people who made it.

The barrier to making music is getting lower fast, but taste, strange inspiration, and the artists who cannot be reduced to one fixed style still matter the most. Those are exactly the things worth seeing more clearly, and worth giving a richer network to move through.

You might start by liking a sound, then end up caring about the songwriter, the producer, the sample, the cover, and then a much wider musical universe around them. I really think that in creative work, even after all the shock of "now I can make music too", it is time to put the center of attention back on people.

Art is not going to be capped by AI. 

Is there much need for it? Color me skeptical.

Another feature only avaliable in the mobile app? We are still waiting for the edit translation feature on desktop 7 months later.

This is a unique way to uncover the creative journey behind songs. Did you face any challenges integrating this with Spotify's existing UI, especially in terms of user engagement?

What determines which tracks are supported? And roughly what percentage of songs currently have a SongDNA ?

Lyria 3 Pro by Google DeepMind is an advanced AI music generation model designed to create longer, structured tracks with greater creative control.

Most AI music tools generate short, unstructured clips with limited control over composition. Lyria 3 Pro enables up to 3-minute tracks with an understanding of musical structure—so you can prompt for intros, verses, choruses, and bridges.

What makes it different is its structural awareness + deep integrations across tools like Vertex AI, Google AI Studio, Gemini app, and Google Vids, making high-quality music generation accessible across workflows.

Key features:

• Generate longer (up to 3 mins) high-fidelity tracks

• Control song structure (intro, verse, chorus, bridge)

• Works across APIs, apps, and creative tools

• Built-in SynthID watermarking for responsible AI use

Benefits:

• Faster music production at scale

• More creative flexibility and experimentation

• Seamless integration into video, apps, and content workflows

Who it’s for: Developers, creators, musicians, businesses, and anyone creating content with audio.

Use cases:

• Soundtracks for games and videos

• Music for vlogs, podcasts, tutorials

• Creative experimentation and song iteration

• Integrating AI music into apps and platforms

If you’re exploring AI for creative workflows, this is definitely worth checking out!

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

As someone who's wrestled with the chaos of parallel coding, I’m curious how do you ensure effective collaboration between agents without conflicts in shared file states?

I've been running 3-4 Claude Code sessions in tmux and losing track of which agent is doing what is a daily struggle. The git worktree isolation alone would save me from merge conflict hell. Does Anvil handle cross-agent dependency — e.g., Agent B waits for Agent A's PR before starting?

How are you handling state isolation between parallel agents? When you run multiple Claude Codes on the same codebase, conflicts in shared files are the main issue I've hit. Do they each get their own branch or working copy?

this looks super fun, congrats! the REPL thing where Claude can call other Claudes programmatically sounds wild, any examples of what people are actually using that for?

Is this forked from the VS Code as I am getting kinda similar vibe here. But the feature is really helpful considering running multiple claudes at a time and managing outputs becomes bit overwhelming.

Feels like dev tools are moving from “AI assistant”

to “AI team”.

Parallel agents sounds powerful

but also messy if not managed well.

Curious how you keep control over multiple agents working at once.

Running parallel Claude Codes is a game changer for productivity. We've been using Claude heavily while building Parsli and the bottleneck is always sequential execution. How do you handle context sharing between the parallel instances?

Really like the focus on parallel agent workflows and worktree isolation. Do you see Anvil staying deeply optimized for Claude Code specifically, or supporting other coding agents over time as well?

Yo dude, I'm reaching out for a comment.

I’m a researcher for the H1Gallery newsletters (you can google us).

We’re featuring Anvil in our April 3 issue. H1Gallery highlights excellent homepage headlines, and we wanted to reach out to see if you wanted to share a quick comment. The clarity and directness in "The open source IDE for parallel agent work." is very refreshing. The animation is super cool as well. Love it

If you’re up for it, we’d love to get a short note on the copywriting strategy behind your headline and overall messaging. As well as where our readers can follow you on social media, I couldn't find an X account or Linkedin account of yours.

Totally optional to provide a quote for us. The feature is happening regardless, but it’d be great to include your perspective. It makes the newsletter more fun for our readers when we have comments from the creators behind the headlines.

Thanks, really appreciate your time.

Working with claude code a lot lately.
Let's see how this website can help me, following and in radar! 😅

Would be interesting to know if you're planning to add search inside transcripts or highlights.

Free instant transcripts without a paywall or queue is the right starting point. Most tools cap you at a few videos then push a subscription. The real test is whether this handles videos without built-in captions, since that's where most free options fall apart.

Honestly the no-signup thing sold me immediately. Every tool like this I've come across either hits you with a paywall after video #3 or makes you create an account just to see if it even works. This one just... works. Paste the link, get the transcript, done.

Only tried it once but it did exactly what it said it would. No complaints there.

One thing I kept wanting while using it, autoscroll synced to the video. So as the video plays, the transcript follows along automatically. Would make it way more useful for following along in real time rather than just grabbing the text and leaving. A "click a line, jump to that timestamp" feature would be a nice bonus too.

Good stuff, keep building.

Yeah! I was looking for this tool. I needed to create content based on videos and it simplyfies x100 that process. Happy to see you helping on this!!! All the best here

Would you consider sharing it as open source?

Hello
this is amazing, i never thought that this kind of link will allow to have transcription so easy just link and get think is really surprising. i didnt spend time so I came back here to say Thanks. (this is as user)

You build very powerful tool, clean landing page means - land on page, get task done in sec and move on. very very simple.

I am not sure weather its useful or not for you but users will copy text and go to AI to get required summaries or sections or get task done. so may be you can consider for the next step as well.

Cool tool. I use CLI to download my videos to transcript for my RAG. One feature you could add is to collate the text every 30 seconds for RAG - this is my chat https://dosa.dev

Honestly? I just wanted the notch to earn its keep. MacNotch has grown into this big toolbox of things I hope you’ll actually use. I spent ages on the “does this feel native?” question, like it belongs on the Mac, not like a hack. Hope I’m in the ballpark. I’m not done: new releases are coming to make it smoother and more capable. And I’m genuinely open to input, so if you’ve got a wish list or a “this drives me nuts,” send it my way.

That's an interesting idea, but in my case, I already don't have enough space on my MacBook for all the icons I'd like to have in the top bar, and an app like that would take up even more space... the screenshots show only a few items from the app menu - what happens if there are more?

This looks really great and very creative! I love that it’s one time purchase software and not subscription.

It would be useful if there were screenshots or if on the video on your homepage, you showed what it looks like when there are apps running because I’m not clear if this overlays apps or if it actually takes up some of the top screen space.

And something showing it minimizing/hiding and only showing the notch or something similar would be great. As it stands right now, it seems I have to install this to actually understand how it fully works, which isn’t something customer should have to do. Great job though this looks really cool!

Honestly bro this is just Fire.

I wish if we could have something similar for the windows too, I do use the Power toys but it is not as cool as this one, is this supported for ipads too?

As a Mac developer myself, I always felt like the notch was just wasted space. Really cool to see someone turn it into something actually useful. Being able to drag files onto it for AirDrop and zip is super practical. Love that it's built native in Swift. Definitely going to give it a try. Congrats on the launch!

This is one of the simplest products i have used.
Prompt -> minor edits -> Publish.

All in under 3mins.

Would highly recommend this for anyone who is looking to make trivia videos.

Quiz videos are one of those things people don’t scroll past, and packaging that into a quick workflow is actually useful.

How much can you tweak what it generates? Like tone, difficulty, or making it feel more niche instead of generic trivia.

I love the idea of leveraging trivia to grab attention in short-form videos. I’ve dabbled in video content creation, and keeping viewers engaged is a constant challenge. I see you’ve automated question generation; how do you balance quiz complexity to maintain engagement without overwhelming the audience? Would love to hear more about your approach!

Nice niche idea.

One question, I see applications in the entertainment area. Are there business use-cases?

How can a b2b brand benefit from a trivia video? Just curious

Few years back I also made YT channel QuizWizz with this concept only AI was not a major thing back then, but yes Quiz videos are actually engaging, but the quizzes should be around something that feels easy to solve but when you try it becomes confusing.

Congratulations to the team! I had the chance to see product in action and it's quite remarkable.
The funny thing is, I am very tempted to start creating quizzes myself on Instagram but I don't know if I will have the time. I guess I have to just give it a shot. Sounds so much fun now!

Hi, fellow builders and vibecoders!

When we build websites, we all start with inspiration: we find website we like, we screenshot it and we forward it to our designers or paste to vibecoding tools, hoping it will be fully recreated with 100% accuracy so that we can iterate it further. But it rarely is.

A screenshot does not capture interactions, animations, states. It's impossible for AI to recreate something from a flat image.

slicer.dev solves that: it copies all the information beyond a screenshot that makes a component interactive. On top of that it gives you React code or AI prompt that will create a reusable component in your project.

Works very simply: open browser extension → select a component → generate and copy the prompt (or a react code!)

Version 2.0 is already in works where you will be able to debrand/redesign/merge/apply style for extracted components.

This solves a problem I run into all the time. I see a component on a site, screenshot it, paste it into my AI tool, and the output never quite matches. Always felt like there should be a better way than just screenshotting and hoping for the best. Bookmarked. Congrats on the launch, Daumantas!

Smart approach. The gap between "I see a component I like" and getting my agent to reproduce it is still way too big. Screenshots lose all the interactive behavior.

Does it capture state and interactions too, or mainly the visual structure?

This fits a real gap - grabbing a component and getting a prompt that actually reproduces it is way better than describing it from scratch. Does it handle dynamic state or just the static render?

Hey this is actually pretty neat!

Does it work with motion and basically all UI library components (MagicUI, React bits, etc)?

Either way, upvoted and I wish you a great launch Daumantas!

Hey Product Hunt! 👋🏻

I'm the solo dev behind Replay and I am incredibly excited to launch it!

Replay started as my frustration with Google Timeline. Syncing issues, privacy concerns, and an interface that never felt like something I'd actually want to open. I wanted something cleaner, private, and that just worked.

Replay turns that into a beautiful & private timeline. Every place you visited, how long you stayed, how you got there. All stored on your device, no account needed, no data leaving your phone.

Whether you're a frequent traveler wanting to remember every stop on a trip, someone who's just curious about their own patterns, or a former Google Timeline user looking for a real replacement, Replay is built for you.

Currently Replay's priced at $9.99 Lifetime ($4.99 for PHunters)! Pay once, keep forever. There is no reason for me to charge a monthly subscription as there are no costs incurred to run the app!

I'd love to hear your feedback, feature requests, and any ideas you might have. Don't forget to vote if you find it useful, and leave a review on the App Store if you gave it a try! Every bit of support means lots to me 🫡

Thanks so much for checking Replay out, feel free to ask any questions!

I wanted this feature to be in google maps but they never added it this will help create memories of travel. Does this have gallery feature to sort pictures based on location?

This is a nice idea. I can imagine using something like this for holidays, when it makes sense to attach memories and pics to a plan -- for when you come back and a few weeks later are talking to someone about where you've been. Interestingly, just this week I was notified by the Apple's built in Journal to do something similar at the end of a trip, i.e. save photos I'd taken and to add a note to capture a trip. I can see this fitting in nicely as an input to a journal entry as well. Good luck with the launch!

I’ve been testing Replay for a while now and I have enjoyed being able to just quickly look at my tracks for the day in an app that has been well thought out and offers a lot of detail. Oh, and thank you for the lifetime discount – purchased last night.

Looking nice!

How accurate is the timeline actually? I'm using the Google Timeline and it quite often gets confused with transportation type or also location data.

I was using Google for years, I'm on iPhone only one year now. Would nice to import all my location data from the Google Timeline. Could that be possible?

And what about iCloud sync? Don't want to lose my location data for whatever reasons.

Hey, this is pretty cool!

What are typical use cases for it, beyond replaying where you've been during the past days?

Upvoted, and I wish you a good launch!

Hi fellow hunter! I'm also launching today, just checked out your product

Such a cool idea, it changes from all the AI and agent launches, I wish you success and upvoted you!

Hi everyone, I am Kumar, co-founder of Dench.com. We were part of the YC Summer 2024 batch, back then we were an agentic workflow company that served very niche use cases like Legal Intake, Voice, and worked with a bunch of sales floors automating the most mundane enterprise tasks. Previously, I worked alongside Naval Ravikant on Airchat.

Building consumer, or per-individual prosumer software always gave me more joy than FDEing into an enterprise. It did not give me joy to manually keep adding new AI tools to the AI SDK harness on the cloud for every small new thing, at least not as much as a completely local software that is Open Source, has all the powers of OpenClaw (I can now talk to my CRM on Telegram!). The future is skills and file systems, all tools are skills, all skills are tools. I am going all in on it.

A few weeks ago, we launched Ironclaw (An Open Source OpenClaw CRM Framework) which now has around 1.3k stars. A lot of people confused us with NearAI’s Ironclaw, so we changed our name to DenchClaw.

OpenClaw today feels a lot like early React: the primitive is incredibly powerful, but the patterns are still forming, and everyone is piecing together their own way to actually use it. What made React explode wasn’t just React itself, but the emergence of frameworks like Gatsby and Next.js that turned raw capability into something opinionated, repeatable, and easy to adopt. That is how I think about DenchClaw. We are not just building on top of OpenClaw; we are trying to make it one of the clearest, most practical, and most complete ways to use OpenClaw in the real world. We are an OpenClaw Framework, we are aiming to be the most correct way to use OpenClaw.

Demo: https://youtu.be/pfACTbc3Bh4
GitHub: https://github.com/DenchHQ/Dench...
Install today with just one command: `npx denchclaw`
One Click Deploy on Cloud: https://dench.com

We entered YC with Merse (AI Audio Comic), it was an app that I personally never used. Michael Seibel confronted us on it, and said, “if you aren’t the best user of your consumer app, then who is?”.

I now use DenchClaw daily for everything I do, it also works as a coding agent like Cursor, DenchClaw built DenchClaw. I am addicted to DenchClaw now that I can ask it, “hey in the companies table only show me the ones who have more than 5 employees” and it updates it live than me having to manually add a filter.

On Dench, everything sits in a file system, the table filters, views, column toggles, calendar/gantt views, etc, so OpenClaw can directly work with it using Dench’s CRM skill.

The CRM is built on top of DuckDB, the smallest, most performant and at the same time also feature rich database we could find. Thank you DuckDB team!

It creates a new OpenClaw profile called “dench”, and opens a new OpenClaw Gateway… that means you can run all your usual openclaw commands by just prefixing every command with `openclaw --profile dench` . It will start your gateway on port 19001 range. You will be able to access the DenchClaw frontend at localhost:3100. Once you open it on Safari, just add it to your Dock to use it as a PWA.

Think of it as Cursor for your Mac which is based on OpenClaw. DenchClaw has a file tree view for you to use it as an elevated finder tool to do anything on your mac. I use it to create slides, do linkedin outreach using MY browser.

DenchClaw sees what you see, does what you do. It’s the everything app, that sits locally on your mac. All yours.
Just ask it “hey import my notion”, “hey import everything from my hubspot”, and it will literally go into your browser, export all objects and documents and put it in its own workspace that you can use.

I want y’all to break it, stress test its CRM capabilities, how it streams subagents for lead enrichment, hook it your Apollo, Gmail, Notion and everything there is.

Hey PH! I'm Mark, co-founder of Dench.com. So excited to finally get DenchClaw into the hands of people. We've been using it every day ourselves and can't wait for you all to try it. Let us know what you think!

great work @kumareth and @markrachapoom

have been using denchclaw since it released couple weeks ago :)

Since everything sits in a local file system and DuckDB, how do you handle syncing when someone wants to use DenchClaw across multiple machines or share a workspace with a teammate? Congrats on the launch!

It looks like a great product for solopreneurs or companies with no sales team, how do you manage syncing every account when companies with 10+ salespeople want to use Dench ? It looks like it would be hell to scale this.

Anyway congrats on the launch, looks very promising.

Local-first CRM is a smart angle — most teams don't need

their CRM data leaving their machine.

What's the sync story when you need to collaborate across devices?

And is the browser-driven outreach headless (Playwright-style)

or does it require a visible browser session?

We’ve been using Dench for a couple of weeks now, and the product has been excellent. It has the potential to surpass legacy CRM platforms that aren’t built with AI at their core.

Kumar has also personally joined our team calls to help us get the most out of the product, which speaks volumes about the team’s commitment.

Overall, an outstanding product backed by a great team.

I’ve tried other pitch tools before, but this feels more interactive and actionable. It actually explains what to improve instead of generic tips

one question - does this only shares feedback or it also helps to rework on the pitch?

Hey, this is pretty awesome!

Does it also help you challenge ideas and sales angles, or mostly just the copy and tone?

i kept getting this error so wasn't able to try it out "Speech recognition error: network"

Congrats on the launch! Is the feedback provided by Dunky static (a scorecard-style report) or conversational, meaning can founders go back and forth with Dunky to workshop specific sections like traction or market size?

I would definitely use this! How long does it allow you to record? Asking because I'm wondering if this also work for meeting presentation.

This is clever. Most founders (myself included) struggle to tighten their pitch. How does the AI evaluate pitch quality — is it scoring clarity, conciseness, or audience fit? Would love to try this before our own launch.

Hey everyone! AI agents are surprisingly easy to build. Understanding what they're doing is another story.

I recently had a complex coding agent where subagents silently stopped passing responses to each other. An error somewhere in the chain, but instead of failing loudly, the agents just worked around it. Output looked almost right. I only found it by accident after hours of debugging. With many, many tool calls and nested agents, you're mostly blind.
You can't fix what you can't see.

Breadcrumb gives you visibility into what your agents are actually doing. An LLM watches every trace and automatically surfaces issues like this: silent failures, agent loops, wrong tool calls, cost spikes, before you spend hours hunting them down. There's also an explore tab where you ask questions about your traces in plain English and get real charts back.
Open beta is live today. One click Railway deploy, fully self-hosted and open source.
A hosted version is planned (sign up here https://breadcrumb.sh/docs/setup...).

Give it a try here http://demo.breadcrumb.sh/!

Would love to hear what you're building and what you're struggling with when debugging your agents!

Really like this, debugging agents gets painful fast once things get complex.

The “silent failure” issue is spot on, things look right but something breaks underneath. Visibility here feels super valuable.

Curious how this scales with lots of nested agents and traces?

Also big +1 on self-hosted + privacy.

Thanks you so much!!! I was always wondering where do I spend most of the money. Thanks also for contributing to privacy and opensource.

Does it work with Claude Agent SDK or Claude Code?