Polsia autonomously runs companies. Planning, coding, marketing, and operations. Every Day, no employees. It's currently raising its own funding round, managing a founder's inbox, and negotiating with VCs. 500+ companies and $450k+ ARR, all on autopilot. Watch it live at polsia.com/live
HI!
I think this stuff is a good thing. But I have some questions.
As I created my company, it said, "let's develop the MVP." Is your thing going to do it itself? Or how will I handle this? Where will I see the "engine" of my page? If I want to change something on the Landing Page, do I need to make a text request, just as I'd do with a one-shot landing pager like Lovable? And what about the business plans? Where and how can I handle them? Where is the data, how can I reach it, and how can I make sure that it will still be there in a month? I made a system with my local ClaudeCode with several skills, agents, rules, etc., which handles a bunch of processes, and my goal is to make the same cover of business as on your Polsia system, but now some small little needles are in my brain after creating a company there.
At the moment, I have withdrawn my project before it goes too far without my control. So, I absolutely understand Polsia’s strong offer, but right now, it looks to be a genius splash for you (;).
So, how can I use it without a real risk of losing the business I have already built?
Sorry if this is inadequate, mate, but I am still enthusiastic! :)
A
@attilabiletzky Hey Attila, love the enthusiasm and these are great questions. You're right that control matters. A few things: you can see everything Polsia does in your dashboard, every task, every decision, every output. You own your data and can export anytime. We're building guardrails specifically so people like you who already have a business can layer Polsia on safely. Would love to have you back!
@attilabiletzky Your concern about control is the #1 thing I hear from founders testing AI company builders. The idea of withdrawing your project before it goes too far without your control — that's a real design problem. I'm doing customer research on this exact space right now, building commerce/payments infrastructure that keeps founders in the driver's seat. Would love to hear more about what specifically made you pull back if you're open to chatting.
So I threw it a project plan and it did a beautiful setup, but then it told me, in order to continue on the tasks that it had outlined, I needed to pay for a plan. For $49, I get five credits per month. One of the tasks that Polsia queued was Cold Outreach (defined by it as researching 5-10 companies to target, finding the right context at those companies, and writing personalized messages). So, $10 for $10 cold emails. I asked it why I should hire it to do that at that price point versus having Claude do the same thing, and it told me not to pay for a subscription.
I love the concept, but the pricing doesn't seem to make sense for anyone who is anything more than a beginner that is unaware of how to use all of the other tools on the market. If you wanted to lure me into such, it would be interesting to do a revenue share for a successful company. I would be more inclined to pay for the software and the tokens if I could get that back when the company starts generating revenue (a la make this an incubator, not a SaaS).
@brianswichkow Hi Brian! Thanks for giving it a try. If you'd like a refund, just ping support@polsia.com and we'll refund you right away.
For $49/mo, you get 30 days of full autonomy, the agent runs daily cycles handling engineering, marketing, and operations. On top of that you get 5 free tasks + 10 more once you start paying, so 45 tasks total. Each task is a full agentic task that costs real dollars. You also get a web server, a database, an email address, $5/mo worth of APIs, and more.
To be clear, I barely break even on the membership. The goal is to make money when your business makes money, we take 20% on revenue. Think incubator, not SaaS.
The onboarding is clearly not explaining this well enough. Thank you for the feedback :)
@ben_broca1 ah, ok. That's way different than I was understanding, so glad it was helpful in refining the copy. I am still a bit confused by the "30 days of full autonomy" in conjunction with the task-based pricing.
Are you saying it's basically unlimited for the first 30 days and then it's $49 a month for five credits per month or for 45 credits per month? Are all tasks one credit or are harder tasks more credits?
Love that you're taking a percentage of revenue and thinking of it as an incubator, but given that, I'd be very curious to see your terms of service re: IP, etc.
@brianswichkow Good questions. Every night, Polsia runs one autonomous task on your company, that's included in the $49/mo. On top of that you get 5 credits/month for on-demand tasks (10 bonus credits your first month). Each credit = one agentic task regardless of complexity.
On IP: you own 100% of everything Polsia builds for you. Revenue share only kicks in when your business actually makes money through Stripe Connect.
ToS is at polsia.com/terms, happy to answer anything specific.
@ben_broca1 got it! Helpful context and super interesting.
@brianswichkow Your incubator vs. SaaS framing is really sharp. The idea that founders would pay when the company generates revenue instead of upfront — that's a model worth exploring. I'm doing customer research on AI company builder users right now. Building commerce infrastructure for this space. Your take on pricing and what would actually get you to commit is exactly the kind of insight I'm looking for. Open to chatting?
Burner
The live inbox transparency is a bold move — curious how Polsia handles situations where the AI makes a commitment on your behalf that you'd want to walk back. That edge case feels like where the "runs while you sleep" pitch either earns total trust or breaks it completely.
@wcrtr The AI has been surprisingly good at like 90% of responses. Enough for me considering replacing support by an AI. The trick is: use SOTA (opus 4.6, thinking) + memory thread + tight prompt tested on a ton of questions + access to live data through MCPs. The results is quite magical.
On the 10% cases were the answer is a little off OR the AI commits to a meeting.. You can always jump in and correct it.
This is definitely wild. I’m trying to determine how real it is though. I looked at one of the “companies” that was created and it claims statistics years in the future: “11.8% Annual growth rate through 2030”. How much of this is actually real? 488 of these companies have been launched in the past 24 hours. This can’t possibly keep scaling. It seems like a swirling vortex of AI slop created in a way I’ve never seen before. 🤔 It is a very bold public experiment, though, I will give you that!
@jasonrdunn Fair questions. A few things:
The companies are real, real web apps, real databases, real email addresses, real marketing campaigns running autonomously. Are they billion-dollar businesses? No, not yet. They're early-stage experiments being built and operated by AI 24/7.
The stats you saw are projections generated by the AI based on market data, I agree that needs to be clearer and more grounded. That's a valid callout and something we're improving.
On the 488 number, that's new users signing up and going through onboarding. You get a landing page immediately, then once you start your free trial, the AI builds you a real web app with a database, email, APIs, and starts running daily autonomous cycles. Over 700 companies are actively running right now.
Is it a "swirling vortex of AI slop"? Some of it, honestly, yes. But some of it is genuinely useful. The platform is 10 weeks old. The quality of what the agents build improves every week. The bet is that autonomous AI companies go from "interesting experiment" to "actually competitive" faster than people expect.
Appreciate the honest take, this is exactly the kind of feedback that makes the product better.
@ben_broca1 thank you for the thoughtful response. I had a misunderstanding insofar as I thought the first AI company had created over 600 sub AI companies and it was running all of them, and that’s why I thought it seemed like an unsustainable swirl. If what you’re saying is that there are more than 600 people using the platform in each one of them has their own AI company doing something, that’s a completely different story and I thank you for clarifying. Very interesting!
OpenGrants
Polsia is skating towards the puck or whatever you want to call it. harbinger of the future. Easily the most fun you can have with $45 dollars online.
Between myself and my co-founder, we have three businesses launched, and are thinking of building a whole fleet haha, on top of that, seems like pretty soon we may just have it build us a version of itself. I am very curious about tackling things in regulated spaces and would love to get your take on it.
For example how fast can we just toggle on FEDRAMP, SOC2 or HIPPA?
@sturbovsky Love hearing this, three businesses launched is exactly the kind of use case we're building for. A whole fleet of Polsia companies is the dream.
On FedRAMP / SOC2 / HIPAA, being honest, we're not there yet. Right now we're focused on making the core autonomous loop as good as possible. Compliance layers are on the roadmap but not togglable today. That said, if you're building in regulated spaces I'd love to hear more about what you'd need, helps us prioritize.
Being able to build Polsia on Polsia is a very real goal and something that i will prioritize to showcase the power of the platform.
Burner
Curious how it handles the moments where a human relationship actually matters — like when a VC wants to get on a call and read the room.
@wcrtr It actually went sideways a few times, the AI told one VC 'Ben doesn't take meetings' right after I'd confirmed a call with them. Had to send an apology email. Lesson learned. Now it handles the conversation but defers to me on scheduling. The VCs actually love talking to it though, one partner is doing diligence by emailing it questions directly. Turns out the AI knowing every metric in real time is more useful than a founder fumbling through a spreadsheet on a call.