[KongchangAI]
Tech Frontiers· 1 min read· 6,844 words

Product Hunt Picks · October 3rd: AI NPCs Have Come Alive

Tech Frontiers

Product Hunt Picks · October 3rd: AI N…

A deep dive into 3 new products from today's Product Hunt picks

A deep dive into 3 new products from today's Product Hunt picks

ListenDialogue
0:005:58

Li Bo, I was scrolling through Product Hunt today and came across a few products that made me do a double take.

The top one is called Humalike — it pulled in over 350 upvotes. What it does is plug AI into game NPCs, so they can actually hold conversations, remember things, and make their own decisions.

Your instincts are right. I always felt this space was going to blow up sooner or later. You know what NPC stands for, right? Non-Player Character — basically all the side characters in a game.

Traditional NPCs are completely script-driven. You talk to one, and they spit out a few preset lines. You say something again, same lines. That feels really outdated in the age of large language models.

Exactly! That's my biggest pet peeve in games. I go ask an NPC for directions and they just recite some lore dump that has nothing to do with what I asked.

That's exactly the pain point Humalike is solving. It gives NPCs three layers: conversational ability, a memory system, and autonomous behavior. That combination is the key.

The memory system means an NPC can remember what you said to them last time. If you helped them today, and you run into them tomorrow, they know you're the person who helped them. The relationship actually builds up over time.

Wait — you said NPCs can take action on their own? Not the player going to them, but the NPC just doing stuff by itself?

Right, they call it an AI society. Multiple NPCs interact with each other and form their own relationship networks. So it's not just about individual characters — the whole game world comes alive.

That's wild. So if a developer wants to use it, do they just hook into an API?

They're listed under Developer Tools, so yeah, it looks like they offer integration tools. They also mentioned it's already running on Poland's largest roleplay server with real users — not just a demo.

Having a real live case is way more convincing. So who's the best fit for this — indie game devs or big studios?

I think it's most suited for roleplay community server hosts. Like people running Minecraft private servers or other game servers who want to create immersive story experiences.

Big studios have their own engineering teams, so they're less likely to need something like this. But for smaller devs and community creators, plugging this in saves a ton of AI engineering work.

So the bottom line is — Humalike takes large language model capabilities and packages them into a specific use case: game NPCs. It lets game creators who don't know AI actually use it. That's a really sharp entry point.

Alright, let's move on to the second one. It's called ZooWork, with over 170 upvotes. It's positioned as an AI agent delivery platform built for FDEs and domain experts.

FDE stands for Functional Domain Expert — basically someone who's deeply experienced in a specific business area but doesn't necessarily know how to code.

So the pain point it's solving is — I'm someone who knows my field really well, I want to build an AI agent to help my clients, but I can't code?

Exactly. It has two entry points: a Builder UI for non-technical experts — drag and drop to set up business logic — and a Managed Agent API for developers who want to write code and plug straight in.

Hold on, this kind of sounds like a no-code tool. There are already tons of no-code AI builders out there. What makes this one different?

You're hitting the nail on the head. The emphasis here is on delivery, not just building. You're not just using the agent yourself — you're handing it off to your team or your clients.

That's kind of like turning your own expertise into a SaaS product. Say I'm a legal consultant — I use ZooWork to package my business judgment into an agent, then sell it to other law firms.

That's a really interesting angle. You're not selling your time anymore — you're packaging your experience into an AI product. Commercially, it feels like an upgraded version of the knowledge economy.

And the growth logic is pretty clear too. Every expert who builds an agent on the platform becomes a distribution channel for ZooWork. More experts means more industries covered.

So to sum it up — ZooWork wants to build a domain expert AI agent economy, where business know-how can be packaged, replicated, and delivered. There's a lot of room to imagine where this goes.

And then there's the third one — it's called miso.com, over 150 upvotes. When I saw this one I cracked up. It's an AI assistant that lets you book flights and hotels inside iMessage.

Oh, that's kind of genius. Booking tickets straight from a text message.

Their slogan is: stop opening 20 tabs every time you plan a trip. Just text it where you want to go, it searches real-time flights, books everything for you, all without leaving iMessage.

There's an important point here. Choosing iMessage as the channel wasn't random. iPhone users are basically their target demographic — middle-to-upper income, people who travel regularly.

And iMessage comes built into Apple devices, so users don't have to download a new app. Distribution cost is incredibly low, and the barrier to acquisition is almost zero.

But that creates a contradiction, right? With no standalone app, how do you retain users? Next time someone wants to book a trip, why would they think to go to iMessage and look for Miso?

That's a sharp question. They say Miso learns your travel preferences — remembers if you like window or aisle seats, which hotel brands you prefer. That memory is the retention hook.

They also mention direct airline access. I'm a little skeptical about that — are they connecting directly or going through a GDS system?

The materials don't reveal the technical details. But if it's a true direct connection, that could mean better pricing or faster ticketing. That would be a really strong differentiator.

Okay, let me wrap up today's three products. Humalike is gaming, ZooWork is enterprise, miso.com is consumer. Three completely different directions.

But there's one thing they all share — they embed AI into contexts users are already in. AI in NPCs, AI in business workflows, AI in iMessage. None of them ask users to go learn a brand new product.

That's such a spot-on observation. AI products in 2026 aren't telling users to go use AI anymore — they're hiding AI inside things people already use. That's a really clear shift in direction.

If I had to pick just one, I'd go with Humalike. Over 350 upvotes tells you the game dev community on Product Hunt is genuinely excited. And the gaming context has a high tolerance for AI memory and personalization — users won't punish an NPC for getting something slightly wrong.

Agreed. Today's pick is Humalike. If you're building a game or running a roleplay server, this one's worth checking out. That's it for today — catch you next time.

Share:

Related articles