Manav Majumdar and a few friends were working on building an AI model for their portfolios when they developed their first model that supported UI development. They posted the open source code on Reddit and Hugging Face, and when they woke up the next morning, the post had gone viral.
Now, five months later, Majumdar has built that model to be the foundation of Tesslate, an emerging startup based in Charlotte that is providing intuitive AI tools for full-stack software development.
The core of Tesslate is Tesslate Studio, which allows users to prompt an AI to build full-stack software for them in just a few hours’ time. Without any knowledge of or background in code, anyone—from individuals looking for a fun project to startups in need of a website—can quickly build their own webpage and software.
“We really just want to enable people to be able to create whatever they want,” Majumdar said. “Your ability should not hold you back from getting customers.”
Distinguishing Tesslate in a competitive market
Amid a rapid rise of no-code software development tools such as Lovable and Bubble, Tesslate distinguishes itself from other platforms by allowing users to download their tools locally, as well as by offering AI tools fit for full-stack development.
When companies want to use AI platforms on the internet, they are often tasked with sending requests out to these platforms’ websites each time they prompt an AI. These requests can contain private company data, which Majumdar said raises concerns relating to IP compliance issues.
With Tesslate, Majumdar has crafted a smaller AI model that, due to its size, can be downloaded locally, minimizing concerns of IP compliance. While many AI models are designed to focus on excelling in every area and discipline, Tesslate is primarily honing in on its speciality of software development, which allows for a smaller, more efficient model.
Local models are also cheaper for users to run. AI platforms use tokens—where words, or pieces of words, from users’ prompts are broken down to help AI process information—to function and produce results. Claude Opus 4, for example, costs $75 per million tokens for output. Majumdar said he estimates that Tesslate can run for just $3 per million tokens for output.
“For us, we just have to be good at coding, or we just have to be good at web design, so our models can be smaller, faster, more efficient and more focused for what needs to happen,” Majumdar said.
Another key feature of Tesslate’s AI tools that sets it apart from other no-code builders is its full-stack development capability. Tesslate utilizes a multi-agent system to enable users to easily build and implement corresponding softwares and features, such as versions of a webpage for both desktop and mobile access.
Capturing the no-code tools market
From its early days as a group of friends’ AI model online to its current iteration as a full-fledged startup, Tesslate has remained a free, open-source set of tools—and Majumdar plans to keep it that way.
“I love open source software. I’ve always built in open source,” Majumdar said. “I want to enable people to do what they can without having to go to proprietary software.”
Majumdar said he plans to keep the core features of Tesslate as free, open-source tools, while also finding ways to generate revenue by monetizing the features that are more specialized for enterprise teams.
QUICK BITS
Startup: Tesslate
Founder: Manav Majumdar
Founded: 2025
Team size: 4
Location: Charlotte
Website: www.tesslate.com
Funding: Bootstrapped
The market for no-code tools is currently around $40 billion, Majumdar said, but he anticipates that to grow into a $1 trillion market by 2035. Within the no-code tools industry, Tesslate is a viable option for independent startup founders, mid-level enterprise teams and full enterprise tools.
Tesslate is already working to capture that market through a new partnership with REACH, a startup focused on creator and brand marketing. Through the partnership, Tesslate’s software will power not only REACH’s software stack, but also 100 other companies that REACH is working with.
Currently in its pilot phase, Tesslate is working with enterprise companies to help test and train its AI model so that it can compete with top models. The startup is also working on launching its next set of tools, which will focus on model training and agent development.
Companies interested in trying out Tesslate’s software can join the waitlist now.
