huggingface/aisheets

Build, enrich, and transform datasets using AI models with no code

What it solves

AI Sheets is a no-code tool designed to simplify the process of building, enriching, and transforming datasets using AI models. It removes the technical barrier for users who want to apply large language models (LLMs) to tabular data without writing code.

How it works

The tool provides a spreadsheet-like interface where users can use AI models to generate or modify data in columns. It integrates with the Hugging Face Hub via Inference Providers or can be configured to use local LLMs (such as those run via Ollama) as long as they follow the OpenAI API specification. For larger scale data generation, it includes scripts that can be run via HF Jobs using either the Inference Client or vLLM.

Who it’s for

It is intended for data scientists, researchers, and non-technical users who need to to enrich datasets or generate synthetic data using open-source AI models.

Highlights

  • No-code interface: Build and transform datasets using a spreadsheet UI.
  • Flexible Model Support: Access thousands of open models from the Hugging Face Hub or connect custom local LLMs.
  • Scalable Generation: Support for HF Jobs and vLLM for processing larger datasets.
  • Deployment Options: Can be deployed locally via Docker or pnpm, or hosted on the Hugging Face Hub.
  • Web Search Integration: Optional integration with the Serper API for web-augmented data generation.

Related

  • Project
  • Project
  • Project
  • Project