Dream Builder
A part of the larger DeepPavlov Project, Dream Builder is a low-code/no-code AI assistant development platform focused on enabling people to build AI assistants of their dreams on top of the Large Language Models.

With Dream Builder and basic experience in prompt engineering, anyone can design, publish, and use their own AI assistants as the ultimate tools for personal growth, helping people to achieve their goals and reach their full potential with the help of cutting-edge AI technology.

About DeepPavlov
Dream Builder is a project driven by DeepPavlov. DeepPavlov is a non-profit artificial intelligence research project founded in 2017. Our goal is to enable people around the world to use the powers of the AI assistant technologies to get things done, learn more about the world and themselves, and become better versions of themselves with the help of our AI assistants in different form factors, from phone apps and gaming NPCs to robots. Our team of computational linguists, machine learning and deep learning researchers, software engineers, designers, product managers, and philosophers work tirelessly to achieve this goal, constantly exploring, expanding and discovering what lies beyond the boundaries of what we currently know.

About DeepPavlov Conversational AI Technology Stack
To accomplish our goal, DeepPavlov team has over the years designed and built an open-source conversational AI technology stack called DeepPavlov that facilitates the development of the complex dialog systems. DeepPavlov project is recognized for both its achievements in 2017-2022, as seen on the Conversational AI Industry Landscape Map made by Project Voice, as well as a solid player in the emerging space of the LLM-enabled chatbot development frameworks in 2023, as seen on The Foundation Conversational AI Technologies landscape map made by Cobus Greyling of HumanFirst. DeepPavlov technology stack powers Dream Builder.

About our Fundamental Research
Our fundamental research is focused on expanding language model's context length. Our most recent technical report, Scaling Transformers to 1M Tokens and beyond with RMT, shares our latest work on enabling language models to support up to 2M tokens for a number of tasks.

Dream Builder by DeepPavlov