How we built ManyGPT accidentally: Founder's story

ManyGPT wasn't a plan it accidentally landed in public domain, We were building it for us. Let me tell you the story.

Hello everyone, this is Gehadul Islam from Bangladesh, Father of 13 SaaS product and generating around $50k MRR currently (Mostly silent), ManyGPT is the first public launch.

Our team size is about 10 currently, building products every day. As an Ai first company we give our team access to paid version of almost every ai, At least 3-4 subscriptions per head, and if we calculate the minimum $20 then it was $60-80 per month per person (Claude code is not included), We are okay with that, no problem at all. but suddenly a question arrived into mind, do they really use $20 worth of ai in every site? the answer is no, probably using 1 third of the credits we are paying for. And i asked them to create a platform for us where we can bring our own key from those providers and we can only pay what we use and especially we can have 1 memory for all of the ai we are using. We built it, we tasted it and finally it become the big savior, we found out where we was paying $60 for per seat, now we pay somewhere between $7-20 per person depending on the usage.

Life was good, then in an evening we discussed why not we publish it in product hunt as our very first public SaaS? Every member said yes, Fingers are crossed. But i called off the launch that week, because i really wander what makes it different than others? there's a lot of software like that, I then asked my team to introduce visual central memory for the ai for the very first time, It works like a human brain, which is active always, and take important notes from your every conversations and store in a central storage (which you can see literally), It knows about your projects, what you are working with, about your company, about your competitors, about your style, about your thought process, and improve itself every day and help you with better reasoning. So now, you never need to explain same thing into every ai you use, you can literally switch between providers in mid conversation and the new ai would know exactly like the previous one. They have the same brain.

I was not yet happy with the value we added, Then we introduced "team chat" this is another big problem we have been facing for long time, for example as a founder you do early conversation with an ai and this ai is trained for couple of weeks before it become the right partner who help on decision level and you told your team to work on it, they need to go through the same process. We solved it here, you can now invite a team member in mid conversation and they can have the same memory access for that thread and continue with you. (You can choose either you want to give access to only memory or conversation too)

That's all about the product.

One more thing, We are still working on it to improve the entire product and make an ecosystem, We need some early user to help us in this early stage with their valuable feedback and suggestions, that's why we introduces a Life time deal at very cheap as a gesture