PH热榜 | 2026-05-06

as a solo founder, my 'team' is mostly just me and a handful of agents. keeping the context consistent across all of them is a full-time job. kanwas feels like it could save me hours of 're-explaining' the product vision to my dev and marketing agents. awesome @johancutych

Not sponsored or anything: Been using this thing for a while, and LOVE it :) it helped a lot with getting my thoughts in order and writing great strategic docs :)

@predrag_ristic1 Really like that you’re not trying to force everything into chat bubbles. The industry somehow decided every AI product needs to look like another messaging app and it gets exhausting fast.

One thing I’m curious about though is onboarding for non-technical teams. Engineers usually tolerate messy flexible systems because they understand the power behind them, but operations/marketing/sales teams often need stronger structure.

Have you noticed users naturally understanding how to organize work inside Kanwas, or do people initially create chaos everywhere before finding a workflow? Feels like this kind of product can become insanely powerful or completely overwhelming depending on first-time experience.

In the app and genuinely loving it within first 30 mins. You guys have built something great 💪

the human + agent angle is exactly the gap right now. most teams have context scattered across notion, slack, docs, claude projects, and neither humans nor agents can really use it well. excited to try this. congrats on the launch @johancutych

This actually hits a real pain point. Managing context between multiple agents is messy right now . Having one shared “brain” feels like a big step forward.

Before everything, really nice video!! Always good to see someone stepping it up in terms of effort + production 😎 Second, congrats on the launch!! In the world of AI, something like Kanwas is a breath of fresh air. Will be following closely!

love that it’s fully editable. my biggest fear with "ai brains" is when they hallucinate a decision we never actually made. being able to step in and refine the context keeps the agents from going off the rails. @johancutych

Great work team, BTW, how should I think about this compared to Notion plus Claude, or Obsidian plus Claude Code?

I have tried to make this work with Claude chats, Notion pages, GitHub issues, and random docs. The issue is not creating content, it is keeping context usable after the first session. Kanwas seems pointed at that exact gap.

The idea of treating team knowledge as something “living” instead of static documentation is really interesting.

Feels like the challenge over time is keeping the context actually useful instead of turning into another layer of noise as more humans and agents interact with it.

Currently vibecoding my own app and constantly losing context across chat sessions relevant info gets buried in history and I can't find it again. This is exactly what I've been missing. Great job, guys 🙌

Going to give this a go immediately! amazing work - thank you

This is super cool! Manually trying to keep context between team mates is such a pain at the moment. I can see how this would really fix it. Congrats on the launch!

Great job!

When anyone can build anything, knowing what to build is the most important thing. Kanwas will really help with this 🚀

good one! upvoted!

Not sure exactly how I could implement that at our company, but it will clearly fix a big problem for me.

When I am brainstorming or researching on a certain topic, I end up with a big conversation. I need to scroll vertically on Claude or Gemini. The whiteboard form factor you're proposing here would definitely help me organize my ideas and visualize them afterwards.

Definitely worth an upvote and a complete test from me. :)

Congrats on the launch, btw, and best of luck!

This is for sure worth looking into. What I worry about is the space getting overfilled with content which becomes irrelevant over time. How do you plan to solve this issue?

I built something like this internally for our team - literally called it our Brain. The idea was they can use that context in Claude though and build skills against it, etc.

Is that possible, or can you only execute tasks within Kanwas? I love the idea of a Brain we can all visualise & collaboratively maintain/iterate in here but then use in claude.

Hey product people. I am in charge of the canvas interface of Kanwas.

Kanwas is built with the idea of moving from chats to shared spatial context space, where you, your team, and your agents can work together and build on top of it.

My favourite part is being able to control context in the easy way, starting from scratch, or starting from yesterdays research, or just cross referencing latest positioning angle with the newest competitor updates.

From the start of building it we are using it as team, and collaborating through all the challenges, from working on strategy and gtm, to updating feature specs, doing users insights from the posthog events and user calls, and even doing this product hunt launch.

Happy to see you using it and getting the feedback! Here to answer any question!

Where does Kanwas fit if we use Notion for docs and Linear for tickets?

How opinionated is the agent? Does it mostly organize what is already there, or does it push back on assumptions and ask questions too?

I am glad this is not another closed workspace where everything disappears into a proprietary database. Markdown/YAML plus Git history makes the product much easier to trust.

Hi Product Hunt 👋

Shadow is a real-time AI on the call that completes your post-call to-dos while you're still on the call

The problem: A user told us something we couldn't ignore:

"I go into back-to-back calls. I miss my meals. Someone wants a proposal, someone wants metrics, someone wants a follow-up. I hang up, and there's a new to-do list every single time. I feel like a robot just running endlessly."

We did the maths. 5 calls a day. 20 minutes of post-call admin. 5 days a week. That's 8+ hours — one full working day — lost to busywork, every week.

The solution: Shadow was already on the call. It already had the context. So we asked: what if it just handled the to-dos before you hung up?

While you're still talking, Shadow detects action items in real time and executes them. Today, that means:

• 📝 Notion doc creation: notes and summaries written as you speak

• 📅 Follow-up meeting scheduling: booked before you hang up

• 📧 Draft email: written automatically based on what was discussed

Coming soon: Slack, Salesforce, Jira, Google Sheets, PostHog, GitHub, LinkedIn, Twitter and more. This is just the start.

What's new in Shadow 2.0: We've moved from a browser app to a native desktop app.

Shadow now:-

• Auto-detects your meetings the moment they start, no manual setup

• No bot joins your call, invisible to everyone else. It just works.

Who is it for? If you live on calls, sales, consulting, recruiting, ops, Shadow was built for you. Anyone who goes over back to back calls.

One question, If you could automate one thing that happens after every call, what would it be?

Drop it below, it might be the next thing Shadow handles.

Shubham

Co-Founder, Shadow

shadowlabs.ai

Hey Product Hunt 👋

We’re excited to launch Shadow today.

We started building Shadow because we kept seeing the same problem again and again: meetings don’t really end when the call ends.

You finish the conversation, but the actual work is still waiting for you:

• writing follow-ups,

• updating tools,

• creating notes,

• sharing docs,

• assigning next steps, and

• trying to remember everything that was promised.

Most AI meeting tools help capture what happened. That’s useful, but we felt the bigger opportunity was not just capturing action items, but helping complete them while the conversation is still happening.

That’s what Shadow is built for.

Shadow listens during the call, understands the context, and helps complete meeting actions in real time - so by the time the call ends, the follow-ups, updates, and next steps are already moving.

We’re still early, but we strongly believe meetings should lead to finished work, not more work after the meeting.

Would love to hear your thoughts, feedback, and questions.

Hi Product Hunt,

Built Shadow because the real cost of a meeting isn't the meeting, it's the hour of follow-ups afterward. Shadow sits in your calls and handles the action items for you: emails sent, docs updated, tasks created. You just review and approve. Excited to share it with PH!

This sounds incredibly exciting, a strong, well-thought-out solution to a genuinely frustrating problem. It’s amazing to think how seamlessly tasks could get done in no time, without needing constant hands-on effort.

I'm the dev, AMA

Hello Product Hunt 👋

I'm Hersh, founding member at Shadow leading GTM. Wanted to share what we've learned building this.

The pattern:

Over the last six months, I've spoken with hundreds of founders, freelancers, consultants, recruiters, and operators. Different roles, different industries. They all said some version of the same thing.

"The actual call isn't what's killing me. It's everything that comes after."

The follow-up email. The invoice. The doc share. The CRM update. The Notion recap. The next-step Slack message.

A 30-minute call creates 90 minutes of work.

Why this matters now:

We've spent a decade adding tools to capture more of what happens in our calls. Notetakers record it. Notion stores it. Slack notifies the team. Every one of those tools makes you do more after the call ends. Not less.

We started Shadow with one question. What if the work just got done while the call was still happening?

What we built:

Shadow runs on your call, understands what's being asked for, and executes the work in real time.

Someone asks for the doc. The link is in their inbox before they finish the question. Someone asks for the latest numbers. The data is on screen mid-conversation. Someone asks for a follow-up. It's scheduled before you hang up.

By the time you say goodbye, the to-do list is already done.

Who it's for:

Anyone whose day is shaped by meetings. Founders, freelancers, consultants, recruiters, ops leaders, sales teams. If you finish your day with a stack of admin that came from your calls, Shadow is for you.

A question:

If you could automate one thing that happens after every call this week, what would it be? Genuinely curious. The answers might shape what we ship next.

Hersh
Founding GTM

shadowlabs.ai

Can it also help create user stories in productboard automatically? I go over lot of customer calls. I want AI to take my notes and put it in my tools

Can Shadow schedule follow ups and send docs on its own, or is there still a human layer present?

This one feels personal. I have struggled with the pile of follow-ups after almost every call, and it’s honestly the most draining part of meetings for me. Love how clearly you have defined the problem and the direction to solve it. Really excited to see how this evolves.

This is actually super relatable 😅 every call quietly turns into a mini to-do list.
Love the idea of handling things during the call instead of after. Curious how well it picks up on action items in meetings with larger groups (>10 people)

How is this different than the other AI meeting assistants? whats the differentiator? The innovation? The novelty? :D How does it stand out from the 1000s of live meeting HUD AI systems for meetings? Sorry for the roasty question. but you asked for it so here it is 😸

Congrats on the launch, folks! Just looking at it from a marketing angle, I think many people would be looking at Shadow AI and thinking: yet another AI notetaking tool.

What would be useful is knowing what makes Shadow AI different? What is the compelling use case that Granola, Fathom, Otter, Fireflies, Loom, and so many others don't provide?

Zoom notes and Gemini notes include action items in their summaries after the email.

A good start is a comparison matrix for people already using these existing notetaking tools on what Shadow AI can do that these other tools cannot.

Every call creators a to do list is such an accurate framing .Most of the pain starts after the meeting ends.

Love the focus on reducing cognitive overload .Remembering action items across multiple meetings is exhausting even for highly organized teams.

This is a genuine time saver. Especially since the end of a meeting is the time with lowest energy levels and maximum procrastination potential.

This sounds like a magical tool. Every meeting is not just like a meeting, it's a list of to-do lists that you get. And I think if some parts or all parts of it can be taken care of, I think that's fantastic. Congratulations on the launch. :))

every call ends and ur evening is already booked with "send the proposal" and "share those metrics".

most tools just summarize after the fact, shadow doing it live during the call is the right shift. congrats on 2.0 team !!

That “I feel like a robot just running endlessly” line is painfully relatable 😅

Really interesting direction feels like trust and timing will be everything for tools like this.

How does this deal with the language switching mid-conversations if we speak multiple languages in the same conversation?

The idea of an AI-assistant for video calls is great. I tested several tools while looking for the best solution for me and my team, but I still haven't found the one I'd use regularly. The app looks great, although the setup process took some time. I'll test it during an upcoming call. Good luck!

Turning calls into completed work instead of just notes is a big shift.

If execution during the call works reliably, this could be a game changer.

Looks great !
Does it also help me talk to my notes as well. Many a time, i want to come back to my meeting asking relavant questions about chat that happened.

@shubham16180 Congratulations. And happy product launch.

Can have coding agent too ?

Hey everyone!

The pace of shipping software keeps getting faster and it's been so much fun building at the bleeding edge, and trying to help as many people ship products people love as possible. We've had quite the grind to get this one out so we're hoping you all enjoy :)

New Features:

• Remote workspaces - ship code across across devices / team members, continue working from anywhere

• Automations - schedule agents to pick up tasks for you later (think of Openclaw's heartbeat feature)

• Superset CLI - gives agents superpowers, and unlocks powerful new workflows

• MCP refresh - works with v2 and backed by an expanded toolset

• A general app refresh - the same Superset you know and love, just a bit more elevated

I'd call this a push to get out the building blocks for everyone's code factories, there's crazy new workflows that are possible now that even we're still discovering.

Excited to see what you all build, and if you have any feedback, we're here to help!

Hey all,

I'm Kiet, one of the creators of Superset. Originally created Superset as a local-first IDE with the ability to run dozens of coding agents in parallel on your machine. After many months and thousands workspaces created daily from users, we're hitting the physical limit on a single laptop. Our users can have 30+ parallel workspaces on their machine running multiple coding agents each.

That's why we built Superset 2.0. It's a ground-up rewrite of the core Superset engine to allow for users to connect to any remote machine and run coding agents there as if it's their local machine. This enables users to scale to hundreds of agents and be able to run with their agents from anywhere, even when their own laptop is off.

The last few months have seen incredible adoption from the most cutting edge teams from all over the world. I'm excited to see what you will build with Superset!

I was using Cursor, then Claude Code, then CC inside Cursor, and finally switched to Superset and haven’t looked back.

Awesome that I no longer need to worry about disk space limits killing my flow

My go to IDE. Superset has totally replaced all others for me including Cursor and Conductor and is now my default for coding with agents.

Just came here to genuinely support superset.
Have been using them since few weeks now. I don't even touch anything other than superset now.

They've really been a game changer for us.
We had our founding engineer come to SF all the way from Australia. And the first thing we did was got him a Mac so he could use Superset. You guys are genuinely amazing!

nice

I've been using Superset for a while now, and I'm excited to see it here. The team has been shipping a lot, tons of improvements each couple of days. I personally enjoy using it to run parallel agents.

been watching superset since v1 and the jump to remote workspaces is huge. running agents from anywhere even when ur laptop is off is the kind of thing that sounds simple but changes how u actually work. proud of the team for shipping this, congrats kiet and satya

Out of curiosity, who is running 100 parallel agents on real production codebases and what is the workflow that earns its keep? What is the most surprising one you have seen?

Been on Superset almost from day 1. It made my life so much easier when it comes to handling multiple terminals and multiple git worktrees in parallel. Such a game changer!

So looking forward to this v2, in particular remote workspaces 🔥. Keep rocking folks 🚀

Running Superset and love it! Congrats on the launch!

Coding just got 100x cooler!

Congrats on the launch! Looks interesting!

I suppose 100 agents run in different branches? How do you handle merge conflicts after or is it something a human should do?

The interesting part here is not more agents. It is whether teams can trust the orchestration once those agents are spread across machines and repos.

How are you handling permission boundaries, failure recovery and audit history when one agent changes code that another agent depends on?

Been waiting for something that handles parallel agent runs without everything bleeding together. The sandboxed environments are the thing. Curious if it works well over SSH on a cheap VPS.

An issue i face in every orchestrator is the "data modelling" in an application, spawning 10 different agents assume different schema for a data representation and add their own custom fields which leads to duplication, or "AI Slop" when trying to merge. Does superset have any plans to tackle this somehow? Very cool launch btw!

Thanks @fmerian! Hey everyone, I'm Ludo, maker of pay.sh 👋

In collaboration with Google Cloud, we're building the open source real-time payment infrastructure for AI agents so they can discover and pay per call for any API.

We are replacing sign-up pages, API tokens and subscription management with a new intuitive tech stack designed for agents to pay for services and goods.

It will be a long journey, and today we're super excited to ship the following:

• Agent Gateway for Google Cloud: a hosted proxy that allows agents to access major APIs including Gemini, Maps, BigQuery public datasets, and twelve others APIs.

• pay: a command-line tool for consuming this new generation of API endpoints. If you're running some APIs, pay can also be deployed as a proxy. Local first, MIT Licensed.

• pay-skills: a catalog of high quality APIs, community-driven. Think `brew` for OpenAPI specs that can charge agents (PAY.md). MIT Licensed, PRs welcomed. Curated and monitored.

pay.sh is built on top of strong standards (x402, MPP) involving players such as Cloudflare, Stripe, Solana Foundation, Coinbase and many others who understand this direction is inevitable.

To give it a try: brew install pay. Let us know what APIs you want added to pay.sh, or any kind of feedback.

I'll stay on top of your comments!

Ludo

Love this direction - One command line to let your agents pay for any API. No accounts. No keys. No subscriptions needed.

S/O to makers @ludovic @vibhu and the @Solana Foundation for this amazing contribution to the ecosystem.

can't wait to hand this to my agent

The agent native angle is what stands out most .Feels like this isn't just a dev tool but infrastructure for autonomous systems.

Pay-per-request, no API keys, agent-native — this is one of those primitives that quietly unlocks a category of products that just couldn't ship before. The thing I'm most curious about is the failure mode: what happens when an agent gets into a feedback loop and burns through 10,000 micro-payments in a minute? I run a small AI alerts product around prediction markets called PolyMind, and the autonomous-decisions-on-noisy-signals problem is brutal even at fixed cost per call. With pay-per-request you've added a real money knob to the loop. Are you exposing budget caps / circuit breakers at the agent level, or leaving that to the agent runtime? Cool launch either way — congrats.

Very cool, is it based on the x402 protocol / whitepaper ?

can't wait to integrate this into my crypto tax api!

Hey! Product Hunt! 👋
I’m Jonathan Attias, co-founder of GYRO.

Our team spent years building products around AI, automation, and payments. Honestly, the team built something that feels a bit like magic.

We started GYRO after noticing a broken reality:

Billions in flight compensation go unclaimed every year.
Not because people aren’t eligible.
Because the process is exhausting.

Forms, rejections, waiting, support tickets. Most people simply give up.

So we built something different.

Connect your email.
GYRO finds delayed flights, checks eligibility, files claims automatically, and helps users get paid.

No lawyers.
No paperwork.
No back and forth with airlines.

The big shift for us was understanding this:

People don’t want another tool.
They want the outcome.

One moment that made us realize we were onto something:
A user found thousands of euros from old flights they completely forgot about.
Then it kept happening again and again.

Today, hundreds of thousands of flights have already gone through the system.

Guess the name fits.

Didn’t expect anything when I tried this ended up getting money back from a flight I completely forgot about 🤯

If you have old flight emails… just try it. Takes a minute.

Found money is the best money 💸. I spent about a minute going through my old flight emails and ended up with a refund I wasn’t even expecting. If you’ve traveled at all in the last year or two, you might be sitting on some cash. Just try it

I had low expectations but i gave it a try and WOW, it found a flight i totally forgot about. So easy!

I’ve definitely left money on the table because the claims process is very stressful.

As a frequent flyer, this product reimbursed me over $3,500 in the past year.

Can’t wait to see what this year will bring.

Never thought I’d find myself secretly hoping for delayed flights lol

In a quick and easy check with a convenient process,I discovered that I’m eligible for a refund on an old flight that was delayed by several hours. In additional we landed at a different destination, and I found out that I’m also entitled to compensation.

Highly recommend, very user friendly and convenient interface.

If you’ve had delayed/canceled flights, do yourself a favor and check this.

I found out I was eligible and got money back with almost no effort.

low effort, high reward.

Just tried it - super smooth experience.

Connected my email and it quickly surfaced flights I didn’t even think about.

Love how simple the whole process feels.

I just discovered I about $1,000 in claims and submitted a refund through your platform :)
Finally a great way where leveraging AI for personal finance!

Manifesting this kind of surprise in my inbox immediately 😭💸

Congrats on launching! I've definitely had flights delayed or canceled and never bothered to chase the refund because it felt like too much hassle. If this actually digs through my inbox and handles it for me, that's a no-brainer. Going to try it out.

I got money for an old British flight that got delayed. Thanks!

This is one of those why didn't this exist earlier? ideas .The friction in claiming flight compensation is real.

I was waiting for this thing for so long! Let’s go!

I wasn’t expecting much, but decided to give it a shot, and I’m honestly impressed. It uncovered a flight I had completely overlooked, and the whole process was incredibly simple.

Love products that just find money you didn’t know you had.

Gave it a try and actually got compensated for an old flight.

super smooth experience too :)

I am just trying it and it found a flight that you have to go through manually and i have to sign a document on docusign - what's the point if i have to go and find all the info again to fill in the form - this should have been automatic and shoudl require only the signature

Congratulations 🎊

Congrats on the launch @norel_eitan
Quick question, does this work across all airlines globally, or is it limited to specific regions/regulations like EU261?

This is great and I bet it scales in a way too. Why stop at just flight credits. I have credits for free pickleball games in my inbox, reward points for Starbucks that have an expiration date, etc.

Confident this can grow into a much much larger business over time. Love it!

the friction point fanis raised is actually the most interesting product design challenge here. the value prop is zero effort, so any step that breaks that contract feels jarring even if its just a signature. the docusign moment is probably unavoidable for legal reasons on certain claim types but it would land completely differently if the form was pre-filled and the only action was literally one tap to sign. the closer it gets to that the more the "magic" framing holds up.

I'm one of those people that loves taking advantage of rewards. But flight compensation is one I've never known how to claim. Even last week my flight was delayed over 2 hours AFTER we boarded the plane.

I can see this product succeeding because of how opaque the whole process naturally is. When a service improves the customer experience and helps them MAKE money, it is usually a recipe for success. Great work and I will give it a try.

Hi Product Hunt! 👋

I'm Davorin from the Databox team, and today we’re excited to share something we’ve been working on for a while: Custom Integrations.

Every team has a tool that runs a critical part of the business but doesn’t connect to their reporting. So to work with that data, teams export it into spreadsheets, maintain scripts, or rely on engineering to build and fix connections.

These workarounds get the job done, but they take time to set up, break easily, and create ongoing work for the people managing them.

Custom Integrations solves that.

It lets you connect virtually any API to Databox and turn the response into a structured dataset without writing code. Once the connection is set up, the data syncs automatically and becomes part of your existing workflow. You can build metrics, visualize them on dashboards, or analyze them with our AI Analyst.

That opens up a lot of possibilities. You could pull historical S&P 500 data and overlay it on your HubSpot pipeline to find seasonality in your sales cycle. Or connect a niche tool your team depends on, like a partner referral system, and finally see its performance next to the rest of your reporting.

Here’s what this means for your team:

• Work with complete data: Bring in data from tools that weren’t part of your reporting and analyze performance without gaps

• Eliminate ongoing maintenance: Replace manual exports and fragile workflows with automated syncing

• Give your team direct access to data: Anyone can explore performance and get answers on their own, without waiting on an analyst or engineer 

Custom Integrations is built for analysts, RevOps teams, and agencies who want to eliminate reporting gaps. 

We’d love your input 👇

What’s the one tool you wish your reporting stack supported—but doesn’t today?

Thanks for checking it out 🙏

I know that this is a business tool.. but boy did I have some fun with this feature. Since everything has an API these days, you really do get to report on anything and everything. Excited to see how this feature skyrockets.

I tested Custom Integrations against a few APIs we had been unable to connect before - tools with non-standard pagination and OAuth2 flows. All of them connected cleanly and the datasets synced without issues. For anyone who has spent time maintaining workarounds for this kind of thing, the difference in reliability is immediately noticeable.

From a technical standpoint, what makes Custom Integrations solid is what it handles behind the scenes - OAuth2, pagination modes, dynamic date values, time zone handling. Each of these sounds minor until it is the thing that breaks your connection. Getting all of that into a setup any analyst can configure without writing code is the real achievement here.

Congrats to the whole team on this launch. It has been a long road and seeing it out in the world is genuinely satisfying. If you are a data analyst or ops person who has been working around a missing integration for months - this one is for you.

This is so easy a CEO can do it.

When this launched, I spent my saturday morning building an integration with AuthoredUp, the tool I use to publish and analyze my Linkedin posts.

I then built a Claude skill that connnects to the Databox MCP to help me draft my posts. It looks through my past post's performance and analyzes why certain posts perform better than others. Then, it uses that to propose post angles, validate my hooks, draft my outlines, and edit my drafts.

I work across a lot of client accounts, and the recurring problem is always the same - data sitting in a tool that does not connect anywhere, so it never makes it into reporting. Custom Integrations gives us a way to handle that without escalating to engineering or maintaining brittle automations. Build the connection, get the dataset, move on.

We spent a lot of time thinking about what it means for data to actually be accessible - not just technically available, but something any person on the team can use without needing to understand the pipeline behind it. Custom Integrations is built around that. You connect it once, and from that point forward the data is just there for anyone who needs it.

Congrats on the launch? Every BI tool is shipping "AI insights" right now. In 18 months, what's the part of Databox that isn't a commodity?

Super proud to be part of the team behind this. Custom Integrations has shown up in so many customer conversations, usually right after someone lists all the tools they can’t live without and we all collectively sigh 😄
Seeing it come to life like this, and done properly, is a big milestone. Massive shoutout to everyone who brought it over the line.

There is always at least one tool in every team's stack that does not connect to their reporting. Something internal, something niche, something built before anyone thought about APIs. Custom Integrations is the answer to that - not a workaround, not a one-off script, but a connection that syncs automatically and becomes part of your workflow the same way every other integration does.

Congrats on launch! The problem is real, I struggled with it firsthand, while running my previous company. The only thing: I hope you don't rely purely on AI provided calculations, since now and then it tends to be incorrect.

the 'export to spreadsheets' loop is basically a part-time job for revops teams at this point. 😅 every time a client asks for a metric from a niche tool, we end up in a manual sync nightmare. being able to turn an api response into a structured dataset without code is a massive time-saver. @zigapotocnik

Simple idea but high impact .Getting all your data in one place without code is still harder than it should be this moves things forward.

From a sales ops perspective, Custom Integrations are powerful because they unlock every tool in your stack, not just the popular ones. The niche platform your SDRs love, the legacy system finance runs on, the internal tool no vendor will ever build a connector for. All of it can flow into the same dashboards as your CRM and revenue data.

That means one full picture of performance, faster answers for leadership, and trends you can actually see across the whole funnel. If it has an API, it can drive your reporting. Nothing in your stack is off-limits anymore.

Combine this with Genie AI Analyst and the possibilities are unlimited.

One of the things that makes Custom Integrations genuinely useful for ongoing reporting is that the datasets sync automatically. You are not managing a one-time import or triggering a refresh manually. The data stays current, which means the reports built on top of it stay current too. That is the difference between a connection teams keep and one they quietly stop using.

I always see how custom integrations are a huge headache for any developer to deal with. This will be interesting to see how it can be used for better integrations around accounting workflows.

the long tail of tools nobody integrates is the actual pain. coming from analytics myself, ive lost full afternoons stitching csvs from some niche tool nobody else uses just to get one chart. being able to point at any api and get a clean dataset out the other side is the kind of unsexy infra work that quietly saves teams hours every week.

Congrats on the launch !!!

Hey PH! 👋

Mapping founder networks is one of those things that sounds simple but is actually a massive pain. You're bouncing between LinkedIn, Crunchbase, and Google trying to piece together who came out of a company, what they built, and how they're connected and you still end up with an incomplete picture.

Alumni Finder does it in one shot. Enter any company and it instantly maps the entire founder alumni network as a visual graph.

Here's what that looks like in practice:

Type in Stripe and you'll see Daniela Amodei (Anthropic), Anurag Goel (Render), Jenn Knight (AgentSync) - all former Stripe employees who went on to raise hundreds of millions. You can see who overlapped with who, which teams they came from, and whether their connection is strong (same team, same time) or partial (same company, different era).

The graph shows:

🌐 Full founder alumni network - every founder that spun out of a company or university
🔗 Connection strength - strong overlap (same team, same tenure) vs. partial overlap (same company, different time)
💰 Funding raised - so you know who built something real
🏆 Cross-company benchmarks - compare which orgs produce the strongest founder networks side by side

A few ways people are using it:

VCs use it to source deals from specific talent pipelines - type in Palantir or Stripe and see every founder that came out, how much they've raised, and whether they overlap with your existing portfolio.

Sales teams use it to map warm intro paths into target accounts - find a former colleague who now works at the company you're trying to get into.

Founders use it to find co-founders - see who from your old company has already made the leap into startups.

Angels and scouts use it to get ahead of deal flow - spot emerging founders from strong talent pipelines before they're on everyone's radar.

Corp dev teams use it to track where talent from acquired companies landed and what they built next.

And a lot of people just use it for fun too by finding who in their alma mater have become founders!

Check it out here: https://tools.crustdata.com
Let us know if you have any feedback!

Seeing spin outs visually mapped is powerful. That's how a lot of great companies actually emerge.

I remember back then when I was choosing schools, we were doing that with completely random information from random websites. As a founder I would have loved to compare schools by the amount that has the most active alumni community and of founders and the one that produced more founders

the paypal mafia and stripe alumni stuff is talked about constantly but nobody could actually map it without spending a weekend on linkedin and crunchbase. turning that into a live graph with overlap strength and funding context is genuinely useful, not just a cool visualization. excited to see what people surface with it.

cool stuff !!

this is amazing. as someone who supports startup founders, this is really helpful. i think it will also be great for creating content about founders.

whoa this is amazing!

Reminds me a little of another startup called Interlink but their use case is helping (mostly young) candidates recruit better by minimizing time spent in the Handshake>LinkedIn>Gmail>Networking Call>Interview>Job twister.

I like the use case here and definitely think it could be helpful to map university networks for younger college grads.

Love the time-to-value. Some of the results are inaccurate though. How fresh is the data?

Are we also able to find the emails and phone contacts of them??

Hey Product Hunters! 👋

I'm Ben, Cofounder of Woz.

A few months ago the world changed. Claude Code with Opus 4.5 made coding agents more powerful than ever before.

We loved the productivity gains, but when we dug into the session logs, we found something painful: half of our tokens were being burned on erroneous reads and re-reads after edits. Paying a premium for productivity is one thing. Paying a premium for wasted tokens is incredibly frustrating.

So we went under the hood of Claude Code and found ways to make it dramatically more efficient.

Today we're launching WOZCODE, a Claude Code plugin that makes it:

Up to 55% cheaper

Up to 40% faster 

+11 points higher on Terminal Bench 2.0

Full breakdown of how it works → wozcode.com/how-it-works

What this means for you as a Claude Code user: 

→ More output before hitting usage limits 

→ Significant savings on API costs 

→ Faster completion of tasks (power users are adding 300+ extra gent coding hours per month!)

How to use it: 

WOZCODE is an official Claude Code plugin, approved by Anthropic. It works anywhere you use Claude Code. Installs in two commands, uses your existing Claude subscription, and drops into your current workflow (CLI, Claude Desktop, VS Code, Conductor, Superset, etc.).

🎁 Launch day offer: WOZCODE is free to try, and anyone who creates an account before midnight PT today gets an additional $100 in savings. Get started at https://www.wozcode.com/

Our team is online and happy to answer any questions. Look forward to hearing from you and seeing how much you save!

Claude loves burning your tokens. Now you don’t have to let it!

Hey PH! Brad here, Cofounder and CTO @ Woz.

Couldn’t be more excited for you to see what we’ve been building. People always ask: faster, cheaper, better… sounds too good to be true. So how does it actually work?

At the core we’ve improved the way Claude Code optimizes context and all the benefits are downstream from that.

Instead of relying solely on the default tools inside Claude Code, WOZCODE uses a set of custom tools designed for efficiency and context optimization. For example, in vanilla Claude Code, a simple "find and edit 3 files" takes 12+ calls (3× Glob/Grep + 3× Read + 3× Edit + 3x Verify Read). By the final step, it’s reprocessing all prior output as input tokens.

WOZCODE collapses that to 2 calls (1× Search + 1× batch Edit). Context stays small, so every step that follows is cheaper and faster. Over a 30-prompt session, that compounding effect adds up fast.

Happy to answer any questions you may have. You can also check out https://www.wozcode.com/how-it-works to see benchmarks and more technical detail.

The cost angle is super relevant. Especially once you move beyond experimentation.

I’m currently using Claude Code to build an internal dashboard, and token usage adds up faster than expected.

How are you optimizing for cost without compromising output quality or context depth?

the cost angle is one of the few honest framings out there right now. we run a heavy claude code workload internally and the bill jumps the moment you go from prototyping to actual production loops, so this hit close to home.

two things i would love your take on:

what is the model-routing logic under the hood, is it static rules per task type or something more dynamic that adapts to prompt characteristics? curious because we have found heuristic-based routing starts breaking once heavy tool-calling enters the picture.

and on the 50% number, is that measured like-for-like on the same task graph or an aggregate across mixed workloads? not pushing back, just trying to understand what setup we would need to reproduce something similar internally.

half of every claude code session is just rereads after edits, ive watched it happen in real time and felt the bill for it. fixing it at the tool layer instead of bolting on another wrapper is the right move.

cool stuff guys !!

hey @ben_collins4
been using this on claude max 20x plan, the token optimisation is real. I hit the $100 free tier limit in a few hours though, which is kind of ironic for a tool that saves tokens. couldn't see what the paid tier limits would be before upgrading, so I paused there. would love clearer upgrade transparency and smoother onboarding.
great job, rooting for you!

Must use product. Been using it for the past little while, and I can finally calm my token-nerves.

We recently refactored our entire platform using Wozcode and it saved us weeks. Setup and auth have been a little messy at times, but overall it’s been a game changer for us at Passive. One of the biggest differences is that we hit Claude limits way less often now, which lets us stay in flow and keep building for much longer without constantly resetting context or starting over. The whole development process just feels cleaner, faster, and more structured. Love the product and look forward to seeing it get even better moving forward!

First I really dig the design of the site! Second cutting the cost of CC, sounds like a dream haha, we can ship more! Thanks for a great product!

Love the simplicity of it! Wondering how you are measuring how much the session would have cost without WOZ. Is it an estimate or you can somehow figure it out precisely?

I've always thought that there was a way to more efficiently save credits for AI agents. Will this also help Claude Code keep it's efficiency and learning for preferred results?

but I wonder, why not just use Pi instead of Claude Code in general? Pi coding agent uses barely no tokens, so token aware users should switch their agent harness, no?

Hey Product Hunt! Arya here, Co-Founder & CEO of @Contrario

My team and I have spent the last 18 months building Contrario in the recruiting space. What’s happening now is that SaaS and AI recruiting tools are everywhere, with people claiming they have proprietary tech that can help you find the most “cracked” talent.  

What people do not tell you is that even the best recruiting tools can’t guarantee that strong talent will reply and interview at your company. That is just not how it works. After working with 200+ businesses and going through YC, what we learned is that the real way to win talent is the best expert recruiters working for you — curating targeted outreach messages, doubling down on what's working, and selling the company's vision in every interview.

So we built Contrario to do exactly that. Check us out: https://www.contrario.ai/  

You import your open jobs, and Contrario will: 

• generate a curated JD with ideal profiles, companies, and more

• find and assign expert recruiters who already made similar hires

• surface high-quality talent interested in working at your company

• schedule interviews for you based on your availability

• and learn from feedback until we find the right candidate

So instead of a slow manual recruiting agency or paying for first-gen AI tools, you get an agentic platform that does everything for you.

A few outcomes: 

• An AI marketing startup hired their core engineering team of 4 in 15 days.

• A people data company hired 2 AEs and 2 operations people in <1 month.

• And 150+ candidates hired in <12 months since starting Contrario including @Wispr Flow

The future of recruiting is humans and agents working together as one. Contrario is leading this wave. 

Who is Contrario for?

Founders making hires. Engineering managers expanding the team. Heads of talent and in-house recruiters looking to grow the team. Anyone who has to hire but doesn't want manual processes and a lack of high-quality candidates to be a bottleneck.

🎁 For the PH community:

Try Contrario today, and if it clicks, get 20% off your first hire, no matter the role.

Get started by booking a demo: https://www.contrario.ai/book-a-demo and select “Product Hunt”

What’s the next urgent hire you need to make? 

Hey! CTO & co-founder here 👋

Really proud of what our team built here at Contrario. Excited to launch today and would love any feedback from the PH community!

Does Contrario integrate with an ATS too, or is Slack basically the whole interface?

As a recruiter, I tend to be skeptical. However, I can tell you that working with the Contrario team the past year has been great. Hard working, innovative, passionate about clients and candidates.

Other companies promise that they will address your hiring with volume and pure breadth of recruitment coverage. Contrario focuses on applying the right matched recruiters with needs resulting is a fast, yet personal experience.

Proud of what the team has accomplished.

congrats on the launch. the network-of-recruiters-plus-AI-agents framing actually makes a lot of sense given how broken the candidate side feels right now.

we see this from the other angle, helping candidates parse and respond to recruiter outreach. so much of what comes through still reads like a template that did not bother checking the resume. the human-in-loop piece you describe sounds like exactly the part that makes calibration possible.

two questions if you have a sec:

what does the feedback loop look like between an agent and the recruiter when a candidate does not fit but is close on one dimension. does the agent learn from the override or wait for an explicit retrain?

and how do you handle the cold-start problem for a startup hiring its first five engineers, where there is no past hire data to calibrate against?

We work with the Contrario team at Listen and appreciate the human recruiters + AI agents approach to recruiting. We've seen great candidates come through the platform for our growth team. Curious which types of roles have been the strongest fit so far across customers?

Is this only for tech roles or can it work for ops, sales, and GTM hires too?

How does this process feel different/same for the recruited?

Cool stuff, how much time does it take human recruiters to perform a background check and verify the Recruiter Agent was not manipulated or lied into hiring a candidate?

fantastic team and product! Contrario is a cornerstone partner that has brought us key team members @arya_marwaha

Love that! Do you plug directly into slack?

the honest framing about tools not being able to make great talent reply to u is the part most ai recruiting pitches skip. pairing expert recruiters with agents instead of trying to fully automate the human bit is a way more believable angle. hell ya !!

Congrats on the launch! On the candidate side, do the rejections come with feedback or help the recruiters generate a feedback? This could help both sides to get good matches.

And on the recruiting side, do you plan to have integration to apps such as Granola to ingest notes into feedback?

the curated jd feature sounds lowkey useful. writing job descriptions is the one task i always procrastinate on. does the agent suggest which companies to poach from or do we provide the list?

Do hiring managers get to see recruiter ratings or track records before they're assigned?

this is super cool! need to give this a spin for kanwas!! btw love the slick design of landing page, you guys rock

Hi everyone!

ChatGPT ads are moving from a limited, high-touch pilot into something much more like a real self-serve ad platform.

OpenAI is now rolling out Ads Manager Beta for US advertisers, where businesses can register, add billing, set budgets and bids, upload ad creatives, create campaigns and ad groups, launch ads, and monitor performance. In other words, it starts to look much closer to the ad platforms marketers already know.

The buying side is getting more complete too. CPC bidding now sits alongside CPM, with reporting for impressions, clicks, spend, CTR, average CPC / CPM, and conversions when measurement is set up.

One route is monetizing natural attention. Another is charging for productivity. ChatGPT ads make that split much more visible. Maybe one of these becomes the classic AI business model, or maybe the real answer is still being invented?

I'm particularly interested in ChatGPT's ad display formats, ad placement sizes, their impact on user experience, and when it will be available to global users. I haven't seen any ads while using ChatGPT, only screenshots on online media.

Oh boy, sad that it is only available in the US now 😢

Really curious to see how this thing will convert...

Advertising inside LLMs feels like the next frontier, but it’s definitely a hard problem to solve without breaking the user’s flow. I’m currently building an inline translator (Fenly), so I spend a lot of time thinking about non-intrusive UI in chat environments.

My main question is about the user experience: How do you ensure the ads feel like helpful suggestions rather than just "noise" in the middle of a conversation? Also, how are you guys handling the attribution side—is it easy to track if a chat actually led to a conversion?

Really curious to see where this goes. Congrats on the launch!

Software engineers today are struggling with the expectations of their organizations that they leverage coding agents to move at vibe coding speeds while still maintaining the production quality required to solve enterprise-sized problems.

In the old days, the motto was "features, quality and schedule: pick two." Today organizations expect that agentic engineering deliver all three. And with software factories and Gas City, we can do exactly that for the first time in software engineering history.

Myself and Julian are the OSS stewards of the Gaslandia ecosystem, including Beads, Gas Town and Gas City.

For me, Gas City represents the pinnacle of 25 years as a product leader building tools and ecosystems for software engineers.

Julian is the chief architect of and contributor to Gas City, building the software factory that itself runs Gas City at a velocity heretofore unknown in the world of software projects of any kind.

Really like how this release treats AI coding agents as an orchestrated production system instead of isolated tools, making “software factories” feel like an actual operational model rather than a buzzword.

Hi Product Hunt! We're back! I'm Kylan, CEO and co-founder of @Inworld.

Some of you might remember when we launched Inworld TTS here. It went on to become the #1 ranked voice AI on Artificial Analysis, voted best in blind listening tests by thousands of real users. That meant a lot to us, so we went back and rebuilt the model from the ground up.

Today we're launching Realtime TTS 2.0. Try the live speech-to-speech experience at realtime.ai. 

Here's the thing we kept hearing from builders: voice AI was built for audiobooks and voiceovers. It sounds good, but it sounds like a human reading from a script. If you've ever talked to a voice agent and thought "something feels off," that's why. Realtime conversation is a completely different problem, and we decided to solve it.

What can you build with it?

• Companion apps that adapt to your user's mood and tone in real time through natural language voice direction

• Language tutors that switch languages mid-session with the same voice, no re-recording

• Characters that sound exactly how you describe them with text-based voice design

• Support agents that get every code, name, and number right with improved alphanumeric handling and International Phonetic Alphabet (IPA) support

So what actually changed?

Natural conversationality. We trained the model on conversational speech instead of narration. You get natural rhythm, breath, micro-pauses, the cadence humans actually use when they talk to each other. Every voice you build on TTS 2.0 sounds like a person in conversation, not a narrator.

Conversational awareness. TTS 2.0 is informed by the full audio context of the multi-turn exchange. Not just the current sentence, the whole conversation. How it speaks adapts to how it was spoken to. A line delivered after a joke lands differently than the same line after bad news. The model knows the difference because it heard what came before.

Full voice direction. You steer the model with natural language the way you'd direct a voice actor. Not preset emotion tags, full descriptions: "act like you just got home from a long day, tired but warm." Combined with inline controls for specific moments ([whispering], [sigh], [excited]), the voice is as controllable as it is expressive.

Text-based voice design. Describe a voice in plain text, generate it. "A posh british man, aged 30-40, speaking deliberately" Iterate on the prompt until it fits, save it, deploy it. No casting calls, no recording booth.

Crosslingual fluency. One voice across 100+ languages with on-the-fly switching inside a single generation. Your voice identity is preserved across every language. No re-recording, no managing separate voices per locale.

Realtime TTS 1.5 is still #1 on the leaderboard. TTS 2.0 takes that quality and adds everything that was missing to uplevel realtime conversation.

Learn more at inworld.ai/tts.  Happy to answer any questions in the comments.

– Kylan

Hey everyone, Andreas from the Inworld team! I've been pumped about this launch for weeks and I'm so excited that we finally get share TTS-2 with you all. If you want to hear what it can do, jump into the playground at inworld.ai/tts and try voice design or steering for yourself or play with our realtime demo at realtime.ai. Would love to hear your reactions!

#1 TTS just got better!

I'm most excited about the improvements made in cross-lingual. It's so seamless to have an engaging conversation and switch between multiple languages like English, Hindi, then French and it's the same voice.

training on conversation instead of narration is the right call. every voice agent ive tried sounds like an audiobook reading my support ticket back.

congrats team !!

Realtime TTS 2 is our best model yet.

It's designed to be a frontend of a voice interfaced application of any kind and scale.

Besides naturalness and multilingual quality improvements, in this iteration, this model can't be actually called a "yet another" TTS. Because similarly to speech-to-speech models, Realtime TTS 2.0 was trained to be explicitly steered to provide the most appropriate response, given the conversation context and agent's goal.

Check it out!

The voice control seems to be crazy good, you can just describe the tone and it gets really close without all the tweaking. Feels more usable than most TTS tools I’ve tested. I am gonna test it!

So I tried it, Speech to Speech. It confuses itself and hallucinates very quickly with just basic questions and conversation, I asked both bots how are you, what are you doing today, and what are you doing for dinner. Both gave me completely different spectrum of answers. They gave alot of filler responses like hey, hmm, huh, which I can understand why those are there. But Jason started telling me how to increase the gain of my television set, and Sarah thought I was going to a party. Also the vocal fidelity is alot to desire, in speech to speech. Just my honest feedback so far. Keep at it.

The #1 TTS on Artificial Analysis just got a major capability upgrade. Most voice AI hears what you say. Realtime TTS-2 hears how you say it.

Had a great conversation with Myles. Highly recommend trying it yourself at realtime.ai. It's truly impressive.

It sounds too much like audio book narration. I guess it was trained on that input? Same thing that plagues every single elevenlabs voice. The only voice that sounds human out there is the alloy voice from open ai. and thats an old ai voice. its so strange. this field should be wide open. competative. whats going on ? what an I missing?

Hey folks 👋 I'm Cormac, co-founder of DevAlly.

We built DevAlly because accessibility compliance kept falling into the same gap between what legal needed, what engineering could prioritize, and what product teams were actually shipping.

Most testing tools give you a report, whereas DevAlly gives you a workflow.
Simply scan your product, and you'll get a backlog of issues prioritized by severity, with code-level fixes. When you've addressed your issues you can generate VPATs and compliance dashboards in realtime, giving customer procurement teams exactly what they're asking for.

We're a small team out of Ireland and this is our first Product Hunt launch. We would love your honest feedback, especially if you're in QA, engineering, or product development.

Ask us anything 👇

Super psyched be launching DevAlly to the world today! Creating a more accessible world for everyone, one web app or website at a time

So excited to be launching DevAlly today! AI is making it easier to build software than ever. We're here to make sure what you build works for everyone.

Delighted that what we've built is finally out in the open. If you've ever found yourself landed with an accessibility audit and overwhelmed at where to begin, this is what you've been missing ⭐️

Awesome product guys!

Some very clear messaging on the landing page with striking visuals so your web designer is excellent.

I used it to scan my websites homepage and I actually got a near perfect result! Yay.

I didn't get to try using it on my application thought and I'm wondering how does it handle getting passed the authentification step?

And would there be a way to run this against, for example, a temporary test environment website. For example the test-website features for firebase.

accessibility always ends up in that weird limbo where legal flags it, eng deprioritises it, and product just ships. going from a report to an actual prioritized backlog with code level fixes is the part that turns it into work people will actually do.

congrats on the launch guys !!! really cool stuff.

Delighted to be launching DevAlly today! Very proud to have our work out in the open and excited for what's to come!

Massively proud of the team behind this launch. Getting to build alongside people this sharp, on a mission to genuinely improves the web for everyone, is a rare combination. Excited to finally share what we've been working on!

I'm a fan. NYC's line culture is next level.

ten cameras pointed at the most cursed lines in nyc is exactly the kind of weirdly specific project that makes the internet fun.

yolo crunching frame counts to tell me when to roll up to lindustrie is genuinely useful too tbhh..

Only in NYC?? Wish it covered more locations.. Amazing concept though.

I absolutely love when I know about a product before I see it on PH! Keep up the great work, great product and pumped to see how it scales.

I also live right outside of Ceres Pizza and it could use a camera! Its between Prince Street Pizza and and the new L'Industrie location on Grand and gets packed for tickets in the morning

This is gold! It would be a brilliant idea to make it worldwide.

Very very useful! I have always had fascination of wanting to know the queue status at a place before going there.

I had a chance to be on a billboard and appreciate that! :D Have a nice launch! :)

The 350+ template approach for product-in-real-locations is genuinely useful for fast iteration on creative — most teams I've seen still pay agencies for one-off shoots they don't even A/B test. One adjacent angle I keep thinking about: layering narrative/context onto the location rather than just placing product. I built StoryRoute for travel storytelling and the gap I keep bumping into is that "city as background" is solved; "city as a meaningful place" isn't. Are you planning template variants where the location itself becomes the message, not just the backdrop?

Hey! 👋 I'm Ivan, co-founder of Magic.

My team has been deep in AR/XR and 3D graphics for 10+ years. We know this space inside out.

About a year ago we started Magic - honestly, because our users kept asking for the same thing:

"Beautiful videos without the AI weirdness"

No prompts, no guessing, no "why does this person have six fingers." Just open a template, get a great video.

We come from computer graphics, so quality was never negotiable for us. Our approach is pretty different from most AI video tools - we work with real footage and add digital content on top. That's why the output is predictable. Brands need that.

The moment I knew we were onto something: an agency told us they'd just spent $5,000 on a video. They made the same thing in Magic for $1. That kind of story kept coming.

Hundreds of thousands of people have tried it now. Guess the name fits...

The problem wasn't budget, it was consistency' that's the line. That reframe alone is better positioning than most AI creative tools manage in an entire landing page. Brutal question as requested: what does failure look like in the 10% that doesn't work first try? That's where I'd want to stress-test this.

Impressive work! It’s refreshing to see an AI video tool built specifically for e‑commerce rather than general animation. The VFX and packshot templates make my product look like it’s in a real commercial, and the pricing feels fair for small businesses. Would love to see more templates for apparel or packaging in future updates. The concept of drag‑and‑drop product videos is exciting, but the site felt cluttered and I had trouble scrolling through categories. I also couldn’t locate any clear privacy or data‑use statements. Before I’d feel comfortable uploading my product shots, can you clarify how images are stored and who owns the generated content? Better onboarding materials would go a long way. Keep it up :-)

real footage with ai layered on top sidesteps the uncanny ai look and the copyright headaches in one move.

smart angle honestly !

The "system + AI = revenue stream" framing is sharper than most launches articulate it. The pattern that makes this work isn't actually the AI — it's that someone with strong opinions packaged a defensible system, and the AI just removes friction on top. I see this same dynamic with niche knowledge channels: people who quietly build genuinely useful expertise on YouTube end up converting better than the polished generalist channels (I run a small modeling channel on YouTube: https://www.youtube.com/@Mod3Loop and the highest-converting videos are also the ugliest). Curious how Once UI is thinking about lock-in vs. portability — if I build a $10k engagement on top of Magic Studio, what's the story when I want to fork the system?

Hey Lorant!

Really clean work on this. Love how polished everything feels without being overcomplicated. Definitely looks like something that actually speeds up building instead of adding more layers.

Hey Product Hunt 👋

We built Spotit because we kept watching people - smart people - get stuck on apps they have been using for years. Hunting through menus. Googling "how do I crop in Figma" for the fifth time. Asking ChatGPT and getting back a wall of text describing buttons they couldn't find.

Spotit lives next to your cursor on macOS.

Press ⌃⌥ anywhere. Ask a question in plain English. Spotit points at the exact next thing to click, on your actual screen. for one tap answers that's it. For longer tasks, it walks you through every step.

You do all the clicking yourself. That's the point. Most AI tools want to do the work for you. Spotit shows you how, so you actually learn it. Next time, you won't need AI for that thing.

It's a Mac app, available today. Free to try.

Three things I'd love your help with:
1. Try it in the app you know least well — that's where it shines
2. Tell me what app you wished it worked better in
3. Roast the onboarding if it's bad

I'll be here all day answering everything.

— David

quick question on the privacy side is it processing the screen locally or is my ui being sent to a server? working on some sensitive client stuff so gotta be sure. clean ui btw. @davidtzuke

"YC Application" - @gustaf Check it out! We missed the deadline!

Loved working and using this project, I hope you guys will have fun using it!

-Ciprian

i'll be waiting for the expansion, i'm not a Mac user. Onboarding seems fine. Maybe more definitive content

Hey Product Hunt! 👋

Alessandro here, CEO and Co-Founder of Dynasty.

We’re excited to launch Dynasty: a platform that helps founders exit tax-free using Trusts.

We help founders plan, create, and administer trusts to stack their QSBS exemption (up to $15M+ in potential tax-free gains) at a fraction of the traditional cost, allowing them to take advantage of the greatest tax strategy in the history of startups: QSBS Trust Stacking.

• • •

A little context on why we built this:

We were the founding team at Carta (first sales hire, first PM, first engineer, and first exec). We helped scale the business from $0 ARR to $300M+, and eventually had a life-changing equity outcome.

But when we sold, we paid the maximum capital gains tax (23.8% federal + 13% California).

After we left, we learned something that completely changed our perspective: an angel investor friend who cashed out $40M paid $0 in taxes (perfectly legally) by using QSBS trust stacking. They’d set up trusts for their family and gifted portions of their QSBS shares into those trusts early, multiplying their exemption.

Each trust got up to $15M in tax-free gains could be used as trust-funds for the beneficiaries, and as investment vehicles to compound wealth.

We couldn’t believe it. The strategy existed, it was legal, and it worked, but it was basically locked behind private-client law firms, opaque processes, and $20–50k+ engagements.

• • •

So we built Dynasty to make it accessible to any founder or early employee.

Dynasty productizes the full workflow: planning → legal structuring → trust creation → compliance → administration → ongoing support, all guided step-by-step through software so you don’t need to understand trust law or tax code.

Even more impactful: trust administration is handled through our professionally licensed trust company, so maintaining these trusts is dramatically cheaper and easier than the traditional route — helping you actually capture the benefit at exit. We are the only venture-backed software company to also be a licensed trust company.

If you’re gearing up for a priced round (or even just curious what you could shield), the earlier you explore this, the better. These structures are simplest and most effective when your shares are still low in value due to gift tax implications.

• • •

🤝 PH DEAL: $300 OFF Sign-Up Our pricing is $1,500/year for up to four trusts. PH founders/early employees who come via this launch will get $300 off there first year. Offer ends May 15th.

• • •

What we’ve done so far:

Since we launched in July 2025, we’ve served over 400 founders, angel investors, and early employees, and created over 1000 trusts for QSBS Stacking alone.

We’re here all day so feel free to ask anything. If you want to learn more, we welcome you to book a time with our team.

Can't wait to save founders billions with you guys! Congrats!

This is actually crazy, didn't know you could just plug Open Finance data straight into an AI agent like this. The built-in governance is what makes this actually usable in the real world. We need something like this for India's AA framework too honestly. How does it handle cases where the data owner needs to manually approve access?

Looks cool! But how about the privacy aspects?

Connecting bank data directly to AI assistants sounds useful… but I’d definitely think twice before linking financial accounts.

👋 Product Hunt,

I'm Gavin, the creator of moar.

Almost from the moment I started using AI tools, I kept hitting the same wall: I'd drop a long document into Claude or ChatGPT and get "file too large" or watch response quality drop because the model was wading through bloated inputs.

When I opened the source files of these docs, I found layers of formatting metadata, repeated headers, empty cells and other clutter the AI had to read but didn't actually need. I was paying for a premium subscription and burning most of my context window on pure noise.

I started doing some digging and came to realize that our documents are formatted and designed for human verbosity, not AI comprehension. And then it hit me: to get better results for a human like me, I needed to first make it better for AI.

So I stopped guessing at what AI tools need from documents and I asked them. moar's optimization algorithms have been co-designed with ChatGPT, Claude, Gemini and other LLMs. Every optimization step tested against what the models actually use versus what they ignore.

Through a series of structured conversations, side-by-side output testing and iterative feedback loops, they told me exactly which formatting metadata is dead weight, which structural cues matter, where bloat costs them attention.

The result is up to 95% token savings with zero loss of meaning for AI tools.

Non-negotiables from day one:
- moar is 100% private by design. No server, no uploads, no logs. Your documents never leave your device.
- moar's AI-native document optimization engine is and always will be free.

Would love your feedback, especially weird files or edge cases I haven't seen yet. Honest critique very welcome.