Practical overview
When it fits
Use PandasAI when you want an LLM-backed, prompt-driven analysis step inside a Python workflow, especially for asking questions about datasets or requesting a chart without writing each analysis instruction manually.
Who it is for
- Python users who want to explore data through natural-language questions.
- Technical analysts looking to reduce manual effort on routine data questions.
- Teams building conversational analysis into workflows that use databases, CSV or Parquet data.
Documented uses
- Ask a question about a dataset, such as calculating average revenue by region.
- Compare information across multiple dataframes, such as identifying the highest-paid employee.
- Request a chart, including a histogram, through a prompt.
- Explore SQL, CSV or Parquet data conversationally.
Requirements & limits
The README specifies Python 3.8 through 3.11 and installation via pip or Poetry. It describes the repository as open source and names the MIT expat license, with a separate license for the ee directory. Its example uses an OpenAI API key through LiteLLM. Managed Cloud and self-hosted Enterprise offerings are mentioned as contact-based offerings.
- This is a Python library to incorporate into a workflow, rather than a standalone hosted chat application.
- The README's working pattern configures an LLM separately through pandasai-litellm; its example uses an OpenAI model and API key.
- SQL, CSV and Parquet are named data formats, while the detailed loading example provided is for CSV.
- The Docker sandbox is an optional, separately installed package used in a distinct setup from ordinary library usage.
- The README identifies a license exception for the ee directory.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Ask a question about a CSV
Set up PandasAI with an LLM and ask a natural-language question about a CSV file.
Before you start
- Python 3.8 through 3.11
- PandasAI and pandasai-litellm installed with pip or Poetry
- A CSV data file
- An OpenAI API key for the README's example model configuration
Steps
Install pandasai and pandasai-litellm.
Configure LiteLLM with an OpenAI model and API key, then set it as PandasAI's LLM.
Load the CSV with pai.read_csv.
Call df.chat with a question about the data and print the response.
Expected outcome
A response to a question about the loaded CSV, such as average revenue by region.
Guide limitations
- The steps describe the README's CSV and LiteLLM example; they do not establish every model or data-source configuration.
- The README specifies Python 3.8 through 3.11.
Use the Docker sandbox for a chat request
Set up the documented Docker sandbox and pass it to a PandasAI chat call.
Before you start
- Python 3.8 through 3.11
- PandasAI, pandasai-litellm and pandasai-docker
- Docker
- An LLM configuration and dataframes for the request
Steps
Install pandasai, pandasai-litellm and pandasai-docker.
Configure the LLM and prepare the dataframes.
Create a DockerSandbox instance and start it.
Pass sandbox=sandbox to pai.chat, then call sandbox.stop when finished.
Expected outcome
A chat request runs with the DockerSandbox supplied to PandasAI.
Guide limitations
- The guide covers the sandbox setup and a demonstrated chat call, not Docker installation or environment-specific configuration.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 15 October 2026
- Original website title
- sinaptik-ai/pandas-ai · Official homepage title · retained in its source language