Our Vision
Many organizations want to be ‘data-driven’, but they often struggle to make the most of their data. A large part of that is the disconnect between those that need to use the data and those that can write code. LLMs can unlock the potential of data and transform the way organizations work, perform, and innovate. While there are still important problems to solve in the application of large language models, VerbaGPT aims to make progress towards this vision.
What is verbagpt?
VerbaGPT is an experimental tool that makes data analytics with large language models (LLMs) easy and accessible in a way that preserves the privacy of your data. You can ask questions of your tabular data in natural language and get answers quickly, without any coding.
VerbaGPT is based on a novel technical design that leverages the code generation ability of large language models (LLMs) in a way where not a single row of the data is sent outside your hardware or is accessed by the LLM. VerbaGPT keeps your data how it should be – private. It does, however, share parts of the database schema that you as the user explicitly allow to be shared.
VerbaGPT works with any comma-separated value (csv) data or Microsoft SQL server on your computer or your organization’s network.
Who is it for?
Generally for those who want to explore the use of LLMs to query databases.
If you are someone who is non-technical but wants to utilize your data – you may be able to develop preliminary insights. If you are someone who is technical – such tools can increase your productivity.