smallpond

license

MIT(OSI)

source

public(deepseek-ai/smallpond)

core features withheld

no

Fully OSI-licensed (MIT), full source public.

• https://github.com/deepseek-ai/smallpond/blob/main/LICENSE recorded 2026-08-13

  MIT License text (Copyright 2025 DeepSeek)

• https://raw.githubusercontent.com/deepseek-ai/smallpond/main/LICENSE recorded 2026-09-29

  LICENSE is the MIT text: "Copyright 2025 DeepSeek" and "Permission is hereby granted, free of charge, to any person obtaining a copy of this software".

• https://raw.githubusercontent.com/deepseek-ai/smallpond/main/README.md recorded 2026-09-29

  README states "This project is licensed under the [MIT License](LICENSE)." and installs with "pip install smallpond"; searched the body for enterprise, paid, license key, hosted and cloud service: none.

PyPI downloads for smallpond are the only usage figure published for this tool.

• https://pypistats.org/api/packages/smallpond/recent recorded 2026-08-13

  244 downloads in the trailing 30 days for smallpond

smallpond distributes DuckDB processing over 3FS, and its authors report a GraySort run that sorted 110.5 TiB in about 30 minutes. The README documents no corpus or model built with the framework.

• https://raw.githubusercontent.com/deepseek-ai/smallpond/main/README.md recorded 2026-09-29

  README describes smallpond as "A lightweight data processing framework built on [DuckDB] and [3FS]." with "Scalable to handle PB-scale datasets" and reports "The benchmark sorted 110.5TiB of data in 30 minutes and 14 seconds" on a 3FS cluster.