Online at: https://parquet-viewer.xiangpeng.systems
Parquet viewer is made possible with the funding from InfluxData.
- Query parquet data with SQL ✅
- Query parquet data with natural language through LLM ✅
- View Parquet metadata ✅
- View Parquet files from anywhere -- local file, S3, or any URLs ✅
- Everything runs in the browser, no server, no external dependencies, just a web page ✅
- Read only you query -- won't download the entire parquet file ✅
-
You can use
?url=to load a file from a url. For example,parquet-viewer.xiangpeng.systems/?url=https://raw.githubusercontent.com/tobilg/public-cloud-provider-ip-ranges/main/data/providers/all.parquetwill load the file from github.parquet-vieweris smart enough to only download the data that is relevant to your query, usually a few KBs, even if the file is large. -
If you have parquet files on remote servers, a useful tip is to use sshfs to mount the folder from the server to your local machine. This allows you to access your remote parquet files without opening additional ports for this tool.
nix run .#cli -- file.parquet
2026-01-09T15:20:13.357327Z INFO parquet_viewer_cli: Serving Posts.parquet on http://0.0.0.0:53703
Serving: file.parquet
Viewer URLs:
http://0.0.0.0:53703/?url=http%3A%2F%2F0.0.0.0%3A53703%2Ffile%2Ffile.parquet
http://localhost:53703/?url=http%3A%2F%2Flocalhost%3A53703%2Ffile%2Ffile.parquet
http://xiangpeng-madison:53703/?url=http%3A%2F%2Fxiangpeng-madison%3A53703%2Ffile%2Ffile.parquet
Press Ctrl+C to stop the server.
It compiles Parquet, Arrow, Datafusion, OpenDAL to WebAssembly and uses it to explore Parquet files, more details.
After you have installed nix, the dev environment is setup with:
direnv allowWhich takes care of the cross-compiling of Rust, C++, C to WebAssembly, and everything else you need to run the project.
dx serve --profile debug-strip
dx bundle --releasewasm-pack test --headless --firefoxnix build .#webOutput will be in result/ directory.
nix build .#vscode-extensionOutput will be in result/ directory.
nix build .#docker
docker load < result
docker run -p 8080:80 parquet-viewer:0.1.31If you use Parquet Viewer in your research or wish to refer to it in your publications, please use the following BibTeX entry:
@software{parquet_viewer,
author = {Xiangpeng Hao and Jigao Luo and LLM and {ParquetViewer developers}},
title = {Parquet Viewer: Interactive Parquet File Explorer},
year = {2025},
url = {https://github.com/XiangpengHao/parquet-viewer},
note = {Online at: https://parquet-viewer.xiangpeng.systems}
}Licensed under either of
- Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)
at your option.
Be aware that most of the code is generated by AI, resistance is futile.
