New year, new Mann-E?

In the previous post we’ve discussed the new way of Mann-E, our breakthrough generative artificial intelligence platform and we mentioned that we value how Web 3.0 may be able to change the world.

Although Web 3.0 was a success, we need to rethink every single aspect of our platform in the year 2025. The year 2025 will be an important year and in this year, we’ll have the AI Agents Revolution and also AI video generation hype. 2023 It was image, 2024 it was huge text models, now we have to go on our way towards new trends to survive.

What happened to us in 2024?

2024 was a great year for Mann-E and our set of products. In this section of the post, we’re going in details about Mann-E and how our different projects made our name a brand.

Mann-E is also open source and we value digital freedoms and rights of our users. The model being open source, grabbed the attention of the businesses and startups. So we made big partnerships as well.

Atelier AI

In Iranian digital marketing space, there was a talk about AI in Personalization and this made Muhammadreza Haghiri, founder and CEO of Mann-E think about what we can do about it here in Mann-E.

This thought and ideation procedure resulted in Atelier AI, which is a platform for AI personalization in Persian speaking community and could make a good reputation for our brand as well.

CEO of Mann-E, Muhammadreza Haghiri
CEO of Mann-E, Muhammadreza Haghiri

Jabir Project

In the summer of 2024, we’ve decided to take our LLM (Large Language Model) side of our projects a little more seriously. This lead to Jabir Project, which is a fine tune on LLaMA 3.1 405B and is now available for free on its website.

Also if you look at Jabir’s API docs, you’ll notice that we also provided an embedding model for people who may be interested in customizing the chatbots and models.

Jabir Project then became the foundation of our side product YouTubeLM, which can summarize youtube videos for you.

Rapidens collaboration

In late days of 2024, another AI company contacted us. They wanted our image models and we made a contract in order to provide the models and also support for those models as well.

The product is now up and called Rapidens. Currently Rapidens supports image generation but soon we’ll add other models as well.

What will happen in 2025?

In 2025, we’re going to work on so many new products and ideas with Mann-E being the centric product. For each product, we’re going to make a difference in AI agents world as well.

But as Alan Kay says, those who are serious about the software, should make their own hardware as well. And I guess this quote says a lot!

A new start for Mann-E on Web 3.0

April 14th 2023 was the day we announced the first release of Mann-E on social media. Back then, it was a really good example of Huffman’s famous quote of if you’re not embarrassed by your first release on the market, you’re too late. When we’re thinking about our first release, we think how bad was it, and how people liked even that ugly website with the worst UX possible.

Well, since it’s been almost two years of us being a player in this huge generative artificial intelligence market, we have made decisions to get to the next level. Since we’re now backed by the big names such as RunPod and Modal.

The whole startup space, is orbiting around one thing and that is the famous question of what is your difference from the competitors? and having huge competitors such as Midjourney or Leonardo, we had to have our specific value propositions. This is where the story begins.

What should be our concerns?

This big pivot has been started since I wrote the article title Privacy Focused AI is all we need in my personal website. I sent it to the people whom I knew and asked them for the feedback on my article. Then one wise man said Why not Mann-E becoming the privacy-focused web 3.0 AI company you’re talking about? and it made my mind sparking.

This was the starting point. Mann-E and people around it made the name and reputation by releasing open source models on huggingface and other platforms. As you may know, open source in general has a strong connection to privacy. This was our strength. we had the open source, we could make people sure about open source models we’re releasing and more importantly, we could make people sure that for using our AI, they don’t need to provide Google or Discord credentials.

At this point, we noticed that some people are using those privacy-focused email providers and then some other ideas came along.

The Web 3.0

Although Bitcoin and the whole crypto currency was a think since 2009, the startup opportunities around these concepts started in mid 2014-2015 with the appearance of Ethereum and smart contracts. This gave us some ideas about how we can make Mann-E a web 3.0 company.

The very first suggestion was an ICO, but since we already have stakeholders, it wasn’t a good option for us. Also we could have faced some legal problems (because we still have no legal professional with us).

The next option which we’re still thinking of is to make credits you get on Mann-E’s platform in form of a utility token. It means you can buy them and make images on our platform or transfer it to other users or even sell it or buy it on a centralized or decentralized crypto currency exchange.

A utility token was a good option and we still have it in mind. But the other idea was the fastest, easiest and most reliable one to start the web 3.0 journey with!

The payment gateway

The third idea was to find a way to make people pay in crypto instead of fiat currencies. Although it may seem a little odd and also may push some of our users out of the picture, but we have decided to go with this payment method as an experiment.

This seemed safer than both utility/governance token and also an ICO. Although we may do both of those in future, what we’re doing now is to provide payments by crypto currencies.

The problem here was we didn’t know how we can have crypto currencies as our payment system. One of the friends suggested we create wallets for our currencies, make people pay to those wallets and then check the transaction with an API.

It was the most secure and reliable way but also could take forever to be implemented for each currency in existence. So we decided to look for crypto payment companies. Meanwhile, I remembered one of my friends from university, was working on a crypto currency payment gateway. So I messaged him and yes, his company was up and running.

The problem solved this easy. We had the payment system, we had the AI, we had the user base, let’s make it from the ground up. This is our story of becoming a web 3.0 based and privacy focused AI company!

New models

In this new release of our platform, we offer three different models. One is Mann-E Dreams which is our flagship Stable Diffusion XL 1.0 based model which is famous for being fast. This is currently like our GPT 4o-mini. The fastest and the cheapest.

The next one is called Mann-E Fast and it is our fastest proprietary model and we found a way to make it faster and more stable than before (on old version of our platform, it wasn’t that fast). This model has results similar to Stable Diffusion 3.0 or FLUX Schnell.

We also have deployed our flagship proprietary model Mann-E Pro on our new release as well. Although it’s the slowest and the most expensive one, but it’s worth a try!

Conclusion

Of course in 2025, we may say new waves of generative AI, as we saw in 2023. This is the time older companies should look for ways to stay a live and we chose the way of web 3.0. Although it may not be quite the right way, or at least in its current form, but we may be able to find out how we can do the best for community and also for ourselves.

Since the new way of payments and finance is very juvenile in nature, what we need is your feedback on our platform. You can test our platform as a guest and also you can make an account and do payments to unlock the unlimited potentials of our platform!

Breakthrough AI Platform