Connect with the Product Hunt community

In my 4 or 5 years of working at startups, I ve noticed that people always have a problem when they are told about changes:

free version jumped to a trial with a subscription or lifetime purchase

subscription and lifetime purchase prices increased

You try a new app and leave its notifications on. Soon the same channel is carrying:

• Updates you need to act on

• Routine progress reports

• Reminders to come back

We're planning to split the editor inside OpenMarkdown out as an open-source (MIT) npm package that anyone can drop into their own app. The app itself stays as it is.

Before we do, we'd love to hear from you:

Containment rate (the share of conversations the bot handles without a human) is the default AI support metric. We think it rewards the wrong thing. A bot can contain a conversation by never letting the customer reach a person.

With customers, we track the outcome instead: was the payment made, the policy renewed, the meeting booked, the issue resolved? One real estate customer cut outreach costs by about 70% and finished a month of calls in under a tenth of the time. That's a number a CFO cares about. Containment isn't.

hey PH Ola here, founder of Specsight, launching today

I ve been building Specsight solo since March, with no employees. just me and a team of AI agents. they handle outreach, call prep, fundraising research (they screened 3,000+ VC firms for me), content, support and a lot of the dev work

Coming from a non-technical background, using AI to code an app felt like magic at first. But I quickly realized that getting a feature to work on the "happy path" is only 20% of the battle handling edge cases broken internet and edge-state bugs is where things get genuinely messy.

You suddenly become the manager of a codebase you barely understand. For other non traditional builders here what was the harshed reality check you hit after launching your first AI-assisted project?

I think a pitch deck is almost like a passport for a startup =)

It needs to contain all the important information about the business, but at the same time, it has to be presented in the right, clear, and aesthetic way.

I m realizing how difficult that balance actually is...

Too much information - nobody wants to read it.
Too little - the business isn't clear.
Too much design - the message gets lost.
Too little - it becomes just another boring document.
So, I have a lot of questions =D

After enough "can I run the tests?" prompts, "can I send this to your customer?" risks becoming another button to clear.

Five months ago I asked here how people keep track of their agents. The replies went well beyond coding. In a later discussion, a builder running customer support agents made a point that stuck: people stop reviewing when checking the draft takes longer than writing the reply themselves.

We're debating this internally and would love outside opinions.
One camp says teams want ready-made roles: an AI SDR, a support rep, a collections specialist. You pick one, connect your CRM, set the rules, and it's working in days.
The other camp says every company's process is different, so teams want to define their own roles, tools and approval points from scratch.
We've ended up offering both, but we're still unsure where most teams start. If you were bringing AI into your sales or support team tomorrow, which would you pick, and why?

A year ago, most developers were using one coding agent. Today, it's increasingly common to switch between Claude Code, Codex, Cursor, OpenCode, and others.

But there's a catch: every agent starts from zero. The context, debugging lessons, and workflows you've built up in one tool rarely carry over to the next.

A friend of mine launched there a few years ago, and they were really happy with the results. At some point, they even considered AppSumo one of their main growth engines.

So now I m wondering how it looks today.

Hey makers

"Why are my emails going to spam?" is a question people now ask ChatGPT instead of Google. Until now, the answer was a generic checklist, because ChatGPT can't read your DNS.

Sorry about this one.

1.1 rebuilt everything behind the inbox. I spent hours testing it, and a few bugs still got through. Thanks to everyone who reported them so quickly. 1.1.1 is the repair, with a few small improvements along the way.

You open a document on a train and the connection drops. Being able to read it is useful. Being able to finish editing it and save the changes is a different promise.

For an app maker, keeping work on the device can mean more to maintain. For the person using it, the tradeoff might be fewer AI features or waiting to share changes with a teammate.

Most AI tools on the Mac work the same way: you open them, paste in context, and explain what you were doing. That s fine until you notice how much of your day is spent re-explaining your own day.

I ve been working on the opposite approach with Winq. It knows where your attention is: which screen you re looking at, what you read, and when you walked away. With that, it can help without being asked:

• You step out of a call for 7 minutes, and when you look back, a summary of what you missed is already there.

• You leave your desk, your Mac locks and your music pauses. When you re back, a card shows where you left off.

• You vaguely remember an article from Tuesday. You describe it, and Winq takes you back to the exact spot.

The tricky part is trust. Something that notices you is only useful if it never feels like it s watching you. Winq keeps the camera on your Mac, never saves frames, and only reaches cloud AI when you ask it something.

A few people here asked how Ilse deals with bigger design systems. So far I talked about one direction: Ilse reads your tokens and components, writes a DESIGN.md, and your agent stops inventing new buttons.

The next release adds the other direction: what's already in your code and should become part of the system.

Open the new Consistency tab and Ilse reads the screen you're on:

• Look-alikes: four links with the same classes pasted by hand, which look like one piece but aren't a component. Ilse groups them and names what's off, like "1 differs in padding".

• Off the system: a DS component overridden with a className, a hand-written copy of a variant, a color outside your tokens, a font or size off the scale.

• Agent review: your own agent reads that screen's code once and adds what rules can't catch.

• Fixes show on the page before anything is sent, the same way the panel works. "Make it a component" turns a group of look-alikes into a note for your agent, with your tokens. Ignore something once and Ilse remembers it for the whole project.

At UnleashX, a voice call isn t done when the customer hangs up. A few things happen right after:

1. The summary is ready instantly. Who called, what they needed, and what was decided, in a few clear lines.

2. Nobody on the team has to listen to a recording to catch up.

3. The CRM updates itself. The contact, call notes, and outcome get logged automatically, so the record is always current.

4. The next step gets triggered. If a demo was booked, the calendar invite and confirmation go out. If the customer asked for a quote, it goes to the right person. If they said call me next week, a follow-up gets scheduled.

5. The customer gets a confirmation. A quick SMS, WhatsApp, or email recapping what was agreed. People love having it in writing.

6. The team gets a clean handoff. If something needs a human, the right person is notified with full context, so they can pick it up without asking the customer to repeat anything.

One of our customers told us the post-call part saved their team more time than the calls themselves, because nobody was writing notes or chasing follow-ups anymore.

Be honest...

When the agent says "done", do you read every line?

Lancetta 0.66.0 is out

Settings now tells you where you are

Hi all, Nick from Storiara here. We build pre-production software for film and TV: upload a script and it breaks it into scenes, cast, locations, props and wardrobe, and you build the stripboard schedule, the budget and the call sheets from that one breakdown.

Today we launched Storiara Studio, a plan for studios and production companies that run more than one production at a time. It adds a private workspace with your own branding, sign-in through your company account, two-step sign-in for every user, and a log of who viewed, printed or exported what. Scripts are never used to train AI. It also keeps Movie Magic compatibility, so you can import a Movie Magic Budgeting or Scheduling file, work on it in Storiara, and export it back.

Building FetchSandbox, one distinction worth naming: a tunnel can connect successfully before any app behavior has been checked.

The managed tunnel polls a cloud relay and forwards to your configured localhost port. Connection setup reports proof_status: not_tested. That reflects the design: connectivity alone produces no receipt. Forwarded requests and responses also travel through that relay, so teams handling sensitive data should factor that in.

Behavior verification is a separate run: configure the expected app behavior under a specific operation, run it, and collect the receipt. Keeping setup results and application test results separate lets engineers inspect each piece of evidence independently.

This post is based on a November 1-2 inspection of the published fetchsandbox-mcp 0.7.30 package, not a new app verification run.

A sales manager using UnleashX shared a moment that made our whole team smile.

A lead came in at 11pm from a company they d been trying to reach for months. Their AI employee replied within a minute, answered questions about pricing and integrations, found out they had a team of 140 and needed to go live before their quarter ended, and booked a call for 10am.

When the sales rep logged in the next morning, the handoff note was waiting: who the lead was, what they d asked, what they cared about most, and what was still open. The rep walked into the call already knowing everything. No so tell me about your company. Just a real conversation from the first minute.

They closed the deal that same week.

FinderGit 0.58.0: back, forward, and the keyboard

Your Mac already knows how to go back. Now FinderGit does too:

Build-in-public update from Sponkly (live translation for chats, calls and meetings in 32 languages).

Apple approved the iPhone app today. With Android already on Google Play and the web app at sponkly.com, Sponkly now runs on every platform I planned for at the start.

What is in the iPhone app:

• Translated chats (end-to-end encrypted)

• Voice and video calls with live translation

• Group meetings where everyone hears their own language

• Face to Face (one phone, two languages) and Record (lectures and talks, written and translated live)

• Website chats from SpeakBot, answered right from the app

The one where you can finally grab an email address out of the header without playing double-click roulette with the sanitize layer. Also: the inbox stops sleeping through Gmail's quiet mood swings.

New

• Copy anything in the header. Hover the subject, the sender, or any individual recipient a tiny copy icon pops up. Click it and the raw value lands on your clipboard (just the address, no "Name <>" baggage).

• Long recipient list? Click the " ". When the "to" line doesn't fit on one row, the native CSS ellipsis is replaced by a clickable one. Tap it and the list wraps open with every recipient, each with its own copy button.

Fixed