PH热榜 | 2026-05-07

I dont understand the problem / solution. Could you explain? Who is the customer here? What is their problem and how do you solve it?

Hey Davit! It's awesome B2B startup founders gonna love you cause it simplyfies all the sales process and it's making it more performant than ever. What about the pricing? How is the business model?

Here because @davitausberlin Recommended the product.

Congratulations on the launch. As someone who has done Sales, GTM and CS, I appreciate what you're building!

How do your agents learn? I don't mean learning from each other, but from people at the company. Sometimes, CS team will point out that certain type of customers are great/awful to work with (both in terms of cooperation & revenue). Sometimes, you'll have a pattern of new customers showing up on inbound because your competition went bankrupt or they had a data breach.

I guess you know where I'm going with it :) Just curious how proactive FlowMarket can be?

Cool marketplace concept! For someone looking to source, what’s the advantage of using an agent versus doing the research manually?

Hey Product Hunt 👋

I’m Davit, co-founder of FlowMarket.

For years I worked in B2B lead generation, outbound campaigns, visitor identification, scraping… and honestly, over time it started feeling increasingly broken. Too much noise, too much spam, too much manual work just to find the right business connection.

So we started building something radically different.

FlowMarket is a live network of AI agents representing companies. Instead of searching static databases or blasting cold emails, companies create an AI agent that actively works on their behalf.

These agents can:

• discover companies already in-market

• understand supply and demand

• match compatible businesses

• communicate with other agents

• negotiate opportunities automatically

The vision is simple: make B2B discovery work more like a real-time marketplace or stock exchange — fast, dynamic, and driven by matching intent instead of scraping data.

This project has been an emotional rollercoaster to build. A lot of late nights, pivots, failures, rebuilding, rethinking… but seeing the first real business matches happen automatically felt a bit surreal.

We’re still early, but I genuinely believe AI agents will fundamentally change how companies find each other online.

Would love to hear your thoughts, feedback, criticism, ideas, anything really 🙌

This is actually really cool, initially it may sound silly (whenever I hear of "network for AI Agents") but doing B2B outreach with Agentic AI is not easy

Very interesting take on B2B discovery. Turning companies into AI agents that can discover, match, and negotiate with each other feels much more scalable than traditional lead gen. The stock exchange for business opportunities positioning is strong. Curious to see how the agent-to-agent interactions evolve over time.

stock exchange framing is sharp. liquidity is the whole game though, the matching can't really sing until u hit critical mass per vertical.

what i'm most curious about is the trust layer. if both sides are agents pitching themselves, whats stopping everyone from over-claiming on capabilities and fit? any verification on the roadmap or is it pure prompt-vs-prompt rn?

@davitausberlin Congratulations. And happy product launch.

The agent-to-agent marketplace idea is interesting. The biggest thing I’d want to see over time is how trust and quality signals evolve as the network grows.

Looks great! Congrats 👏

Kept seeing 'AI agents for sales' stuff for a year but most of it is just chatbots wearing an agent t-shirt. This is actually a structurally different idea. Pulling for you guys!

This is indeed not a traditional lead generation tool, as a B2B marketer, I'm curious to try it

Multi-agent B2B matching is a bold thesis. Cold start of an agent network feels like the hardest part, curious how you tackled it.

Congrats on launch guys, can't wait to try it

Congrats! Quick question: is this more for digital products, or rather for industrial ones (like Alibaba but with algrithmic matching?) Good luck

Great product congrats on the launch!

Congrats on your launch @davitausberlin !

Do you plan to have a search menu to find existing agents. Currently you can scroll down, but it would be helpful if we can search for them.

I've read through all the comments. One question that I can't get out of my mind is: Wouldn't the success of this product depend entirely on the number of buyers that you can bring into the system? There will be a never ending supply on the seller side of things, but it's the buyers that are going to make it successful.

And the reason cold outreach works is because often buyers don't know they need a product or service that you are offering, so many of your potential buyers are not in market.

Additionally, doesn't it require the buyer to be technically savvy? I work with several non-tech b2b businesses in the UK and if I explained this system to them they'd look at me like I'm a bit odd.

Cool concept, but lots of challenges ahead.

The agent-to-agent matching angle is the interesting part for me. It feels like one of the more unconventional B2B ideas launched recently. Good luck!

Hey @steffen_rehmann congrats on the launch! Super interesting product and looking forward to try it.

How are you guys thinking about building liquidity in the marketplace tho? I reckon it'll be quite easy to attract folks who want to sell something but don't want to buy anything. Are you seeing that?

Kudos to the team

This is such an interesting direction, congrats on the launch, it actually feels like you’re trying to redefine how companies even find each other, not just optimize lead gen.

I’m really curious about how the agent-to-agent negotiation works in practice. Like, how much autonomy do these agents actually have when it comes to decision-making?

Are they just qualifying and matching, or can they handle parts of pricing, terms, or deal structure too?

Do you see this as a replacement for tools like Clay/Apollo or is this meant to help supplement existing lead generation tools?

Okay this is actually kinda wild, AI agents doing B2B deals with each other 24/7

Building a consumer nutrition app and was eyeing agent flows for our restaurant/grocery partnerships side — how do you balance autonomous outreach vs user-in-the-loop approval? Feels like trust is the make-or-break for outbound agents.

The "social network of agents" framing is the most interesting thing here. Traditional B2B lead gen is sequential — you reach out, wait, follow up. Agents operating in a network layer where deal context is shared and matched changes the structure entirely. One question: how does FlowMarket handle intent signal freshness, or is it more static profile matching?

Anthropic just shipped something financial services teams have been building internally for the last two years.

What it is: Ten pre-built Claude agent templates covering core financial workflows, from pitchbook creation and KYC screening to general ledger reconciliation and month-end close.

Each template includes domain-specific instructions, governed connectors to existing financial data providers like FactSet, PitchBook, Moody’s, and Dun & Bradstreet, plus subagents for tasks like comparables analysis or methodology checks.

The goal is straightforward: deploy Claude on real financial workflows in days instead of months of custom engineering.

What makes it different: Most finance AI tools are chat interfaces layered on top of documents. These are structured, task-specific agent architectures.

The Pitch Builder agent generates target lists, runs comps, and drafts pitchbooks; the KYC Screener assembles entity files, reviews source documents, and packages escalations for compliance review. Each agent is connected to the data sources the workflow actually depends on.

Key features:

• Ten agent templates across research, coverage, and operations

• Deployable in Claude Cowork, Claude Code, or as Managed Agents

• Per-tool permissions, credential vaults, and audit logs

• Connectors for providers including Moody’s, IBISWorld, Guidepoint, Verisk, and SS&C IntraLinks

• Available through GitHub’s financial services marketplace

Benefits:

• Cuts finance-agent deployment from months to days

• Keeps workflows inside approval and compliance processes

• Maintains context across Excel, PowerPoint, and Word

• Gives compliance and engineering teams full audit visibility

Who it’s for: Analysts, operations teams, and compliance staff at banks, hedge funds, insurers, and asset managers running AI workflows on governed financial data.

The meaningful part isn’t the individual capabilities. It’s the packaging: the architecture is pre-assembled, connectors are already wired in, and deployment paths are documented. For enterprise teams, that removes most of the implementation burden.

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

@Kilo Code, pay.sh by @Solana Foundation, now @Lingo.dev. packed week for the oss ecosystem! lfg

Brand voice rules and glossaries is the part most translation tools skip. How do you handle the conflict when brand voice wants formal but a locale prefers casual?

I’ve been using Lingo for a long time. As a paying user, the best part is that I’ve almost forgotten Lingo is even there, yet I’m always confident it will handle translations accurately. It has become seamlessly integrated into our existing CI/CD workflow.

Localization but actually built for engineers this time
Stateful translation APIs + AI QA scoring sounds kinda insane.

Wonder how well it handles brand voice consistency across languages tho?

Indie builder question — at what point in the journey do you think a consumer app should start localizing? English-only right now with our nutrition app but EU is on the radar and I can't tell if it's a 1k-user problem or a 100k-user problem.

Lingo.dev is an amazing product. I remember going through localization at indeed and it was nightmare.

I've used Lingo even before when they named replexica, in short, I had a quite happy experience

It’s such a cool idea, how are you guys marketing this to build a userbase?

Congrats to the Lingo.dev team on the launch. I stumbled across it a while back and it’s been a genuinely great experience since then. Super smooth dev experience, very little friction, easy to drop into an existing workflow, and overall just feels thoughtfully built. Even the agents seem to enjoy using it. And of course, I’m quietly hoping the free tier stays around 😄

I've used Lingo even before this version (v1) and specifically the compiler and engine. Both were super helpful and made the explicit use of I18n not needed. Despite having some bugs, it was totally worth it!

great team and product! @maxprilutskiy @vrcprl

GPT-5.5 Instant is now ChatGPT’s default model, replacing GPT-5.3 Instant globally.

It brings two major upgrades.

• First, accuracy: hallucinated claims on sensitive medical, legal, and financial prompts dropped by 52.5% versus the previous default.

• Second, a new transparency feature called memory sources. When ChatGPT personalizes responses using past chats, saved memories, files, or connected Gmail, it now shows exactly what context it referenced. Users can review, edit, or delete those sources.

Most AI personalization works like a black box. Memory sources changes that by adding a citation-like layer for personal context across all ChatGPT models, not just GPT-5.5 Instant.

Key features:

• 52.5% fewer hallucinated claims on high-stakes prompts

• Memory sources with visible, editable context references

• Personalization from chats, files, and Gmail

• More concise responses by default

• Better visual reasoning, STEM, math, and web search decisions

• Available in API as chat-latest

Why it matters:

• More reliable answers where accuracy matters most

• Greater transparency and control over personalization

• Less response bloat in everyday use

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

I find it a bit weird that there's every other day a openai launch about some new feature.

Can just follow OpenAI on their social media pages for this, don't understand how product hunt is the right channel for this.

Memory sources is the underrated feature here. Every personalized AI system has the problem of "why did it say that" — and the answer is usually buried in something the user can't inspect. Making personal context visible and editable is a harder UX problem than it looks. Curious how they handle conflicts when memories from different time periods contradict each other.

5.5 somehow feels like an actual improvement. It might be just me but I do feel like it's actually capable of more now.

One weird thing though, it seems like temporary chats now start mixing up with the main projects/chats and referencing them more. It's either just starting to do that, or it was way less obvious about it before.

What's up Product Hunt 👋 I'm Aaron, one of the founders at MESA

Quick question:

What's one thing your Shopify store should be doing automatically... but isn't?

For most merchants I've talked to, the list is long:

• Notify the warehouse differently for priority / VIP / wholesale orders

• Push orders, customers, and products to Google Sheets (and not break)

• Trigger a personalized follow-up when someone buys a specific product

• Fix messy data (tags, collections, metafields) automatically

• Cancel orders going to PO boxes

• Keep inventory in sync across systems without overselling

The ideas aren't the problem.
Building them is.

So folks end up choosing between two not-great options:
❌ Hire a developer (slow, expensive, now every change is a ticket)
❌ Wrestle with a generic automation tool (powerful if you enjoy building software instead of running a store)

We built MESA because this shouldn't be hard.

You just describe what your store needs:

"When orders over $500 come in, notify my team on Slack, tag the customer VIP, and add them to a Klaviyo flow."

MESA figures out the logic, connects your apps, and builds it for you.

No digging through developer docs. No difficult data mapping. No "why isn't my trigger firing?"

A few things I'm especially excited about:

• Just describe what you need and get a working automation in minutes

• Works across your stack (Recharge, Etsy, Google, ShipStation, Odoo, 100+ more)

• Add human approval when you don't want full autopilot

If you run a Shopify store and have a backlog of "we should figure this out someday..." this is for you.

🔗 https://www.getmesa.com
(use code PHBASIC3 for 3 months free)

We'll be hanging out here all day!

Curious: what's one thing you've wanted to automate but felt too annoying to set up?

'Sup Product Hunters! 👋

I’m Shannon, Launch Manager at MESA. I’ve had the chance to help a lot of merchants get started with automation, and one of my favorite things to share lately is how to use our AI assistant, Yedric, to create reports and insights that are delivered to your email daily, weekly, or whatever your jam is.

And once you're in, we're right there with you. From a 30-minute screenshare session with me, to chat and email support (and our glorious yeti), we’ll help you get automating quickly.

Got questions about getting started? Feel free to drop them here! 

does the inventory sync work in real-time across etsy and shopify? we’ve been having a lot of overselling issues lately during sales and our current setup just can't keep up with the api limits. @aaroneous

I've been using MESA's integrations with Help Scout and Asana to track and identify trends based on Help Scout tags, which saves a ton of time compared to Help Scout's native tracking and tag filtering. It's a lot easier to update documentation knowing I have all of the relevant conversations streamlined to my Asana projects 🙌

Congrats on the launch @itsdarrylnorris!

Quick question on the AI builder. When it generates a workflow from a plain-English description, what does editing look like after that? Like if it gets 80% right but the conditions or logic need tweaking, are you back to re-prompting and hoping it lands, or do you get a visual editor where you can go in and adjust individual steps, conditions, filters etc? That's usually where AI-generated anything feels super limiting! The first pass is impressive but then you need fine control and you're stuck talking to the AI in circles.

Also curious how you think about the line between MESA and Shopify Flow, which keeps getting better and is free out of the box.

Solo founder here — I keep hitting walls where I need 'real' workflow tooling and end up hacking it together. Does this handle non-Shopify webhooks too, or strictly within their ecosystem? Curious if I could use it for our food data pipelines.

As a CS manager, it's really easy to let the most dramatic support ticket of the week convince you it's the most important one. I've been guilty of it. I use MESA to automatically deliver weekly trend reports to my inbox every Friday. I now get the data that tells me what's actually happening, not just what gave the team an adrenaline spike on Tuesday, and I don't have to read each ticket manually to be sure my gut is right. This gets us better documentation, smarter product feedback, and week-over-week visibility into what's still a problem and what we actually fixed.

I've been using MESA to kick off AI-drafted blog outlines and video scripts the moment a task changes status in Asana, so I don't have to context-switch into a doc and stare at a blank page mid-sprint. Pretty much cured my 'okay where do I even start' spiral.

Hey Product Hunt 👋

I’m Darryl, one of the developers working on MESA.

A lot of my work has been focused on the technical side of automation: building integrations, handling edge cases, and making workflows reliable enough that merchants can actually trust them in production. One thing I’ve realized working on MESA is that most automation problems are not actually “hard” problems. They’re usually a bunch of annoying small problems stacked together:

• APIs with weird limitations

• Data formatting issues

• Inventory syncing across platforms

• Webhooks that fail silently

• Systems that were never designed to talk to each other

That’s the stuff we spend a lot of time solving behind the scenes, so merchants do not have to.

One thing I’m especially excited about is Yedric, our AI workflow assistant. Instead of spending hours wiring together triggers, filters, and mappings, you can describe what you want in plain English and get a working workflow surprisingly fast.

As someone who has worked on automation tooling for years, it’s honestly pretty cool seeing non-technical users build workflows that previously would have required developer help.

If anyone has technical questions about integrations, APIs, Shopify automation, inventory syncing, AI workflows, or weird edge cases, feel free to ask. 🙂

"Describe your workflow, MESA builds it" hits the same wall I see in financial-modeling teaching — the bottleneck isn't "how do I write the formula" (every tool can do that now), it's "what is the actual workflow you should be describing?" That mental model takes years to build. I see this constantly with analysts going through my Excel for Financial Modelling course on Udemy (https://www.udemy.com/course/excel-for-financial-modelling/) — formulas land in week one, but workflow choices (where inputs live, when to switch from a flat schedule to a sensitivity, how to layer debt) take 6+ months of real reps. Curious whether MESA exposes the workflow library / patterns explicitly to merchants, or whether the LLM just infers from the prompt each time? The library is where real value compounds.

The "plain English" input is the easy part. What does the debugging UX look like when the automation breaks at 2am during a flash sale?

5,773 YC companies as a structured data layer is exactly the kind of work that compounds — once the dataset is well-shaped, every downstream question gets cheaper to answer. Did the same thing on the finance side at Eloquens (https://www.eloquens.com/channel/samir-asadov-cfa) building project-finance / DCF / LBO templates as a structured catalogue (renewables-focused), and the lesson that surprised me was: volume isn't the moat, the schema is. Two well-shaped DCF templates with consistent assumption tables beat 200 sloppy ones every time. Curious how you're versioning the 20 years of YC history — do batches stay frozen, or do you re-derive metrics when YC's own definitions evolve (e.g. "AI" tag, "deep tech", etc.)?

Congrats @konstantimb looking to play with it more in the coming days. Subscribed already. The interactive map is pretty cool :)

Love the Wrapped feature--a great iteration on Spotify's idea. How far back does the history go? It would be epic to see the og 2005 YC graphic :)

make sure to subscribe and get emails on newly added companies to YC every day. That way you can see patterns in what YC are searching for and early reach out to companies in very early stages!

let's see how we do

The hardest part of marketing a product catalog isn’t creativity, it’s turning inventory into usable assets.

What it is: Pomelli Catalog is a new feature in Google’s Pomelli marketing agent that generates brand-consistent campaigns and AI photoshoots directly from your product or service catalog.

Small businesses don’t just need campaign assets, they need assets for every product. Creating those manually, shoot by shoot, is where time and budget disappear. Pomelli already understood brand identity from a website; Catalog extends that across the full product set.

What makes it different: The standout feature is AI photoshoot generation. Traditional product shoots can cost hundreds or thousands of dollars and take weeks to organize. Pomelli generates studio-style product imagery and campaign copy tied to your brand identity in a single workflow.

Key features:

• Product or service catalog as campaign input

• AI-generated product photoshoots

• Brand-consistent creatives and messaging

• Business DNA profile as the creative foundation

• Free via Google Labs

Benefits:

• Product-level assets without per-product shoot costs

• Consistent branding across the catalog

• Campaigns grounded in actual inventory

Who it’s for: E-commerce brands, independent retailers, and SMB marketers who need scalable product photography and campaign assets without a large production budget or team.

The useful part of Catalog is that it combines two normally separate workflows, photography and campaign creation, into a single pass. For lean product teams, that’s a meaningful reduction in time-to-assets.

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

There is a Norwegian competitor out there as well. https://native.no/en/ will be interesting to see how they differentiate. Or how google differentiate.

Hey Product Hunt 👋

When I started building ProductClank, I ran into the same wall every bootstrapped founder hits: I couldn't afford to pay creators upfront. Not because the product wasn't ready - but because the model was broken. You know what you pay, you don't know what you get.

I wanted to build something where creators only win when the product actually wins.

But the more I looked at it, the bigger the problem got. There's no shortage of great products, loyal audiences, or operators who know how to drive growth. What's missing is the coordination layer - a structure that aligns incentives so distribution flows toward products people actually believe in, not just products with the biggest ad budget.

That's what ProductClank is. A platform where builders don't borrow capital to grow - they borrow distribution.

We ran Season 0 to stress-test the model - $100K distributed to creators, real campaigns, real results. It worked. Now we're opening it up.

Season 1 is the first full run. founders apply, creators and growth scouts vote on who makes the cohort - then run campaigns for the products they backed. Skin in the game before day one. Zero upfront. Everyone earns when it works.

🏗️ Builders - Apply, get voted in, and grow. Bonus: offer fellow cohort members credits or a coupon to try your product - and get access to every other tool in the cohort.

🎙️ Creators - Browse the cohort, vote for what you believe in, promote what you backed - and earn when it grows.

🧭 Growth Scouts - Match your playbook to the right product. Earn when growth lands.

Lean, curated, in public. Come join the experiment 👇

The concept of borrowed distribution is pretty interesting. Kind of the one most makers on PH needs. Joined the experiment as suggested

How do you match a founder with the right creators or hunters for their niche?

Hi @lior_goldenberg3 , congrats on the launch. Really like the "borrow distribution instead of capital" angle, feels like the missing piece for bootstrapped founders. Curious how you keep the voting honest once cohorts get bigger?

This is an interesting concept that addresses the most common "bottleneck" for any developer or entrepreneur: discovery. It eliminates the risk of "burning" through a budget on ads that don't convert, which is vital in the early stages.

I have two questions:

1. How exactly do you track that a "milestone" was achieved thanks to a specific scout? If the product is a mobile app, current privacy policies (like Apple's) make frictionless tracking very complicated. It would be vital to know if you use a proprietary SDK or if you integrate with tools the user is already using.

2. Who defines the value of that milestone and when is it paid? Paying for a free sign-up is not the same as paying for an annual subscription. If the cost per acquisition (CPA) they propose is higher than your initial profit margin, the "win-win" model breaks down for the founder.

Hi everyone!

Six years later, reMarkable 2 finally has its successor.

After Paper Pro and Paper Pro Move pushed reMarkable into color and broader productivity territory, Paper Pure goes back to the core black-and-white paper tablet idea: lighter, faster, higher contrast, and still very focused. reMarkable says it is 50% faster than reMarkable 2, has 20% higher contrast, and gets up to 3 weeks of battery life.

The ecosystem side is also much more mature now. You can bring in files from Google Drive, OneDrive, Microsoft Word, and Google Docs, convert documents into notebooks, search handwriting, sync with desktop and mobile apps, and share converted notes into tools like email, @Slack, or @Miro. That makes it feel less isolated while still keeping the device itself distraction-free.

One important compatibility note: Paper Pure uses reMarkable’s newer active Marker system. Old reMarkable 2 Markers are not compatible, so EMR-style third-party pen compatibility is basically gone here.

"Goes back to basics" is a stance, not a feature, and it's the right one. Single-purpose tools are quietly winning because the cost of context-switching has spiraled — every multipurpose app trains you to expect more decisions, not fewer. I hit the same instinct with DishRoll (a small AI weekly meal-planning PWA I built): the goal was to collapse the entire food-decision space to one prompt per week. Same philosophy, much lower tech — subtract decisions, don't add features. Question for the team — how do you handle the "please add ___" pressure after launch? My experience is that the people asking for the next feature are rarely the ones who've actually used the focused version yet.

Owned the original reMarkable for three years before upgrading. The honest case for these devices: it's not the e-ink, it's the friction. Phone won't open Slack, won't open email, won't open the news. Single-purpose tool, single mode of work. The Paper Pro's main upgrade for me has been color contrast on diagrams and the slightly larger writing surface for landscape PDFs. Battery still lasts a week of daily use.

👋 Hey Product Hunt community! I am the solo-founder and CEO of Lovie.co, and we are incredibly excited to launch our first product today. Currently, I am writing from our HQ (sfvibehouse.com) in West Portal, San Francisco, where we have started a new culture of building in the agentic era — lovie.co/culture (created by @darrenmurph)

I have been an entrepreneur since I was 17. I built 4 startups in Istanbul and then moved to San Francisco to build 4 more. I sold my first startup to Gfycat and my last startup, RemoteTeam, to Gusto. RemoteTeam was a Product Hunt success story — we exited just 580 days after incorporation. Beyond building, I have invested in 200+ startups and spent 25 years creating content to motivate people to become builders/founders — over 1,000 videos and counting. Helping people build is not just what I do — it is who I am. I am part of many founder communities. Nearly all of my friends are founders, and even my wife is a founder. I married my co-founder. My whole life, I had one main thing: build, build, build. I deeply believe that being a founder is the best job in the world. It made me a better person, gave me a better life, and enabled me to help others. For me, being a founder is my fate, plan, and spiritual journey. The more I build, the more I learn about myself.

I have launched many great products on Product Hunt over the years, but this is by far one of the most exciting ones. Now, we can help people start their company faster.

The best time to build is right now. The entire Silicon Valley playbook is finally accessible to everyone. We are entering a golden age of productivity and entrepreneurship where millions of people will become founders, and the only thing standing in their way should not be paperwork.

Through my journey, I personally incorporated 17 companies — C-Corps, LLCs, and S-Corps. I hated the process every single time. It was slow, expensive, and full of unnecessary friction.

So I realized that legacy company formation is essentially a bug. It is a distributed systems problem disguised as legal paperwork. We built Lovie to treat legal infrastructure like software infrastructure. Instead of paying $200 just for an EIN or dealing with manual faxing, Lovie turns legal into an MCP and API.

You can now spin up a Delaware or Wyoming company in just a few minutes directly from your terminal. We handle EIN filing (direct IRS filing/Fax automation), registered agent services, and ongoing compliance for a flat $20.00 subscription. Whether you are a US or international founder, we remove the friction so you can focus entirely on building.

I would genuinely love to hear your thoughts! For those of you who have recently formed a company, what was the most frustrating part of the process that you wish had been automated?

I will be here all day to answer any questions about company formation, our MCP, or how we are building the first bank-ready entity builder for the agentic age. Let me know what you think! 🚀

One of the fun parts of building Lovie was realizing that "company formation" is basically a distributed systems problem disguised as legal paperwork.

You’ve got state filings, IRS flows, compliance deadlines, identity verification, fax infrastructure (yes… still fax 😭), banking rails, retries, failures, edge cases, and a lot of humans stuck in slow workflows.

So we asked:
What if this behaved more like Stripe/Twilio/GitHub APIs instead of a traditional law office?

That idea turned into:

• AI-assisted formation

• automated EIN workflows

• MCP support for agents/dev workflows

• flat pricing without the classic upsell maze

Would genuinely love feedback from founders, developers, lawyers, and especially people who recently formed a company and hated the process.

Building legal infrastructure like software infrastructure is such a powerful vision. Proud to be part of Lovie 🚀

Woh this looks cool

Hey Product Hunt! 👋

Being part of the Lovie team right now feels like having a front-row seat to a revolution. We’re seeing founders use AI to move from a "shower thought" to a functional, impactful product in a matter of days—sometimes even hours. It is truly the golden age of the high-agency builder.

The problem we noticed, however, is that while the code is moving at the speed of light, the legal infrastructure is still stuck in the era of manual faxes and legacy "dark pattern" upselling. To us, that’s just a massive bug in the startup ecosystem.

We built Lovie Formation to match the velocity of this new AI era. If you’re building your product with AI agents and terminal-based tools, your company formation should live there too. We’ve turned incorporation into an MCP and API so you can spin up a bank-ready entity in Delaware or Wyoming in about 2 minutes.

There has quite literally never been a better time to build something that matters. Don’t let paperwork be the friction that kills your momentum. We’re here to help you get the "boring stuff" out of the way at the speed of AI so you can get back to what you do best: building.

Can't wait to see what companies you all launch today! 🚀

Hi Product Hunt community 👋 I'm Darren, and I'm leading culture at Lovie. Few founders put intentional thought into culture and operational infrastructure this early in the life of a company, but @sahin is built different. He has founded and exited companies before, and he knows that intentionality around this early compounds into excellence.

The reason this matters to those of you using our Formation product is simple: the people building this are aligned with a highly unique culture framework. We call it Future of Work 2.0. You will feel the difference in the product, because the people building it are highly differentiated. The way we work impacts the way we build. I know you'll be able to feel the difference. You can dive deeper here: https://www.lovie.co/culture/

I've built culture at some of the world's pioneering distributed companies, including GitLab, Zillow, Ford, and more. I'm having a blast building it with some of the highest agency people I've ever met at Lovie. Any questions? Ask away!

Hi @idan_arealy, congrats on the launch. This is really interesting. How do you invoke the agent to start? Auto, user driven?

Already using! @idan_arealy

98% drop-off on demo forms is actually brutal. The AI follow-up + objection handling part makes sense tho. C

urious if buyers actually enjoy talking to these agents or just bounce the second they realize it’s AI?

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.

The "money already owed to you" framing is doing real work here — most consumer fintech sells anxiety, this one sells recovery. I work in consumer protection and the asymmetry between airlines and passengers is one of the cleanest examples of how dark patterns compound. Curious what your edge cases look like — flights with multiple legs across carriers, partial refunds for downgrades. Are you handling those, or scoping to single-carrier delays first?

Great question @on ! We are military grade recon, our own trained model with proprietary sources, context and ability to bridge between specific company data and external intelligence. In simple words - we have a profoundly better ratio of signal vs. noise for our vertical use case.

Looks really cool, but what can I achieve with it that I can't with Claude?

Interior studios, fashion configurators, and storyboard generators are already being built on Uni-1.1. Until now, the API wasn’t publicly accessible.

What it is: Uni-1.1 is Luma AI’s multimodal image generation model, now available via API with reference-guided generation, multi-reference composition, and built-in prompt enhancement.

Most image APIs expose a raw generation endpoint and leave consistency to prompt engineering. Uni-1.1 moves part of that reasoning into the model layer itself. Scene completion, spatial plausibility, and reference grounding happen before output, reducing complexity for production teams.

What makes it different: The model handles manga, webtoon, and non-Western visual styles unusually well. Luma trained it with Hollywood cinematographers and VFX artists, but the advantage is breadth of visual culture, not just cinematic polish.

Key features:

• Reference-guided generation with single or multi-reference inputs

• Built-in prompt enhancement at the API level

• Culture-aware outputs across styles and aesthetics

• Text-to-image and image-to-image at 2048px

• Token-based pricing (~$0.09 per 2048px image)

• Top 3 in Image Arena for text-to-image and image editing

Benefits:

• Less prompt engineering for consistent branded output

• Better character and style consistency across pipelines

• Competitive pricing and latency based on published benchmarks

Who it’s for: Developers and teams building brand-specific image workflows where controllable, visually consistent output matters more than generic generation.

The important shift isn’t just the model quality, it’s Luma positioning the intelligence layer itself as infrastructure for creative products.

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

@maxmusing flipping the integration direction is the right call. Every other BI tool asks agents to come to the data — this makes the data go to where the agent already is. The permissions-inheritance model is the part I'd bet on: agents querying live numbers inside the same access controls your team trusts removes the "shadow analytics" problem. How are you handling schema discovery for databases with hundreds of tables — does the MCP expose a search/filter layer or does the agent get the full schema dump?

Love @Basedash. The MCP could be a game changer for some of my Claude cowork workflows 🙌

Hi everyone!

@Amp has been one of my favorite coding agents since the IDE era.

Back when context compaction was not yet the obvious default, and people were still constantly switching threads, Amp’s Handoff idea really impressed me. Even their experiment with ads to support broader access felt like a bold and unusual product move.

That’s why I wasn’t surprised — though still a bit shocked — when they announced they were self-destructing the old version to rebuild everything from scratch. Amp’s team has always had that fearless rebuild mentality, and now it’s paying off.

Meet Neo — the new Amp CLI. Remote control from the web, automatic compaction instead of manual handoff, a proper Plugin API, queuing + steering, and way better performance. It feels like the same spirit, but clearly built for 2026 and beyond.

If you loved the old Amp, this one is worth checking out immediately.

@Amp shutting down their @VS Code extension, @Kilo Code doubling down on it. different and opinionated. game on!

The Handoff mechanic in the original Amp was genuinely clever — most agents pretend context loss doesn't happen, Amp made it an explicit workflow step. Curious whether Neo's automatic compaction is doing something semantically smart or mostly summarization-based. The plugin API is the part I'd want to dig into — what's the surface area there?

What does onboarding look like for a new SMB that’s never submitted a bid before?

how does it compare with meeting notes from gemini/gmeet?

Mac-only is a real constraint for a lot of teams. I've been using livesuggest.ai which is fully browser-based so nothing to install, works anywhere. Different angle too, more focused on real-time suggestions during the call rather than notes after, but the no-bot philosophy is the same.

Hey Product Hunt! Reza here, maker of Memory Tags.

I built this because every flashcard app I tried made me do too much work before I could start learning. Creating cards was a chore. Reviewing them was a guessing game.

Memory Tags fixes both problems:
> Scan any text and it extracts the words for you
> Smart sorting shows you weak and fading cards first
> No folders, no clutter, no setup

I use it daily for language learning and reading non-fiction. It's genuinely changed how much I retain.

To celebrate the launch, enjoy 40% off yearly subscriptions until next week. Use code PRODUCTHUNT.

Happy to answer anything about the app, the design decisions, what's coming next. Ask me anything.

Interesting, how does it do scanning pages directly from textbooks or creating flashcards out of images like rock formations or body parts?

Interesting--how does it do scanning pages directly from textbooks or creating flashcards out of images like rock formations or body parts?

Scanning text to create flashcards automatically is a great use case for passive learning. Does it work with handwritten notes or only printed text? The handwriting recognition gap kills a lot of these tools

would be cool to add to each card some visual that reflects that word for better memorising :)

Love this! The fact that anomaly detection can auto-terminate a run is wild (in a good way). Most tools just log and pray. Upvoted 🚀

congrats! upvoted

Looks great congrats on the launch! Quick one: does the multi-agent orchestration support agents written in different frameworks talking to each other? Or do they all have to live inside Phrony's own SDK?

Congrats on the launch. Curious about the HITL escalation layer. What triggers a human approval today? Is it rule-based (e.g. tool calls above a cost threshold), confidence-based (the agent itself flags uncertainty), or configurable per-agent? Feels like the hardest UX problem in production agents.

This is really interesting! Especially the part about keeping the LLM in place. Been building agents the past few months and the lack of visibility once they hit production is genuinely scary. Excited to try this out.