Grants
We support the most critical open source projects, using an open, data-driven model co-developed with our donor community. Our approach is to complement the OSS funding landscape, sustainably addressing its gaps.
We fund only nonprofit projects without corporate ties or startups behind them. Our ideal grantee looks like a random person in Nebraska.
How We Select OSS for Grants
20+ Data Sources
Our scripts collect data on packages and repos from top ecosystems, including downloads, dependencies, funding signals, etc. This helps us map the global software supply chain.
Community Nominations
Know an OSS project that deserves support? We use nominations to check whether our selection model covers important repos. They don't directly affect ranking for grants.
Nominate a projectOSE Members
Everyone can review our selection model and raise issues in its repo. The OSE Board and Members discuss it on invite-only calls and collaborate to make it better every year.
Open Selection Model
Within each supported ecosystem, the model finds the most valuable repos — the ~1% that carry ~99% of downloads and dependencies. Then it sorts this core by risk metrics, prioritizing the most critical ones. All scores come from open data.
github.com/osendowment/modelDue Diligence
The OSE team contacts top-scoring projects to verify eligibility, independence, and interest in a grant. An operational review ensures funds can be received by maintainers legally.
Distribution
Based on the due diligence results, the Board approves the final list, and grants of up to $5,000 are paid via GitHub Sponsors, Open Collective, or bank wire.
Grantees
August 2026
For our first grant round in 2026, the Endowment had about $11,000 available from investment income generated by the end of 2025. We decided to distribute $10,000 equally among the top projects in the four currently supported ecosystems: Python, Rust, C/C++, JavaScript/TypeScript.
ws
A fast, well-tested WebSocket client and server for Node.js. It powers real-time features in Socket.IO, dev servers, and many web apps. 5.4B downloads on npm in 2025.
github.com/websockets/wslxml
The most feature-rich XML and HTML library for Python, built on libxml2 and libxslt. Scrapers, data pipelines, and document tools rely on it. 1.9B downloads on PyPI in 2025.
github.com/lxml/lxmlSerde
The standard framework to serialize and deserialize Rust data. Almost every Rust program that reads JSON, YAML, or TOML uses it. 333M downloads on crates.io in 2025.
github.com/serde-rs/serdeExpat
A fast, stream-oriented XML parser written in C. Its small size and speed made it the default XML parser on many platforms. Used in Python, Android, macOS, and Firefox.
github.com/libexpat/libexpatws is a popular Node.js WebSocket library with tens of thousands of dependents. It provides a battle tested WebSocket implementation used by several major developer tools and platforms. At the time of writing, it counts over 200 millions weekly downloads.
My first commit to the project dates back to 2015, and over time I became increasingly involved, eventually becoming the main developer and maintainer.
I am grateful to Endowment for their grant to the WebSockets organization, reflecting both on the past and looking forward to the future health of the project and the broader OSS community.
lxml has long been the one go-to tool for all XML needs in the Python ecosystem, reaching from web development to messaging to data science. Its download numbers have only ever grown over its more than 20 years of existence, getting close to 100 million per week these days.
I have been managing the project and implementing major parts of it for most of that time, tuning it to high performance in data processing and bringing it from an experimental side project to a mature developer tool.
20 years ago, making a living as a free software developer required a Real Day Job, but the growing awareness after the Heartbleed incident for our global dependency on OSS and free infrastructure software allowed for new organisations like OSE to appear that help making free software sustainable on its own.