Meta Open Source’s cover photo
Meta Open Source

Meta Open Source

Software Development

Menlo Park, California 9,244 followers

Empowering diverse communities through open source technology.

About us

Meta is committed to the notion that open source not only provides great technology for developers, but also brings the best out in people. Follow us for updates on Meta's open source.

Website
https://opensource.facebook.com/
Industry
Software Development
Company size
10,001+ employees
Headquarters
Menlo Park, California

Updates

  • 🚀 Excited to introduce OpenZL, Meta’s new open source, format-aware compression framework! OpenZL uses a unique “graph model” for compression, enabling modular, high-performance codecs and a universal decoder. It delivers top-tier compression ratios and speed, with minimal CPU usage—perfect for structured data at scale. Built for rapid development and real-world production, OpenZL is already showing impressive results at Meta. Curious to make your data smaller and faster? 🔗 Learn more: https://lnkd.in/gFSJh8E7 📄 Read the whitepaper: https://lnkd.in/gPD7QQyF 💻 Try it out: openzl.org #OpenSource #DataCompression

  • 🔎 Meta Open Source 101🔎 Meet Sapling, Meta’s next-gen source-control system. It’s Git-compatible but designed for massive repos and lightning-fast operations, even at Meta’s scale. Perfect if you need speed and a smooth developer experience. Learn more 👉 https://sapling-scm.com Why Sapling? • Git-compatible CLI & workflows • Blazing performance on huge monorepos • Built-in smart history rewrite & interactive rebase You can adopt it incrementally—use Sapling’s client while keeping your existing Git server. Getting started: 1️⃣ Install Sapling (brew install sapling on macOS) 2️⃣ Clone a repo: sl clone <repo> 3️⃣ Work as usual—commands like sl commit, sl log, and sl pull mirror Git’s feel but are optimized for speed. Want deeper insights—architecture, scaling lessons, and Meta’s journey from Git? Check out the engineering deep dive: https://lnkd.in/g_XnESfi ⚡ Faster commits, friendlier workflows—give Sapling a try today!

  • Hello Python Developers! We’re excited to share that the 2025 Python Type System and Tooling Survey is live - and the typing community would love to hear from you! ✨ You don’t need any prior experience with typing to participate - your perspective as a Python developer is what matters most.💡💻 If you have 10 minutes, please take the survey - your feedback can make a real difference in helping improve Python typing for everyone. 👉 Take the survey here: https://lnkd.in/eMKQHPxU Thank you for your time and insights! #Python #PythonTyping #DeveloperSurvey

  • Thank you, PyOhio, for hosting us this year - we are very grateful for being members of your community! It was an excellent opportunity for us to talk about our new open source project, Pyrefly (https://pyrefly.org/), and the work we have been doing around type-checking 🚀 https://lnkd.in/g3jphVBt

    View organization page for PyOhio

    618 followers

    #PyOhio 2025 was fantastic! We're excited for PyOhio 2026 on July 25-26! PyOhio is able to secure our venue for 2026 thanks to Meta's sponsorship. Meta Open Source This is the first time in years that we've been able to announce dates for the following year at the current year's conference closing events. It was great getting to speak with Danny Yang, Aaron Pollack, and Abby Mitchell from Meta about Pyrefly, (https://lnkd.in/gJfC5PwM). Meta’s support to Open Source, both through code (such as Pyrefly, PyTorch, and Free-Threading CPython: https://opensource.fb.com/) and through fiscal sponsorship to open source and conferences like ours, help keep the community thrive by investing and supporting the long term success of Python and PyOhio. Being able to plan a year in advance makes many elements of a conference significantly easier to manage and it makes it possible to better server our community when we have the financial freedom to handle important expenses. This year we were fortunate enough to be able to accommodate some attendees by providing American Sign Language interpreters, and that was only possible thanks to having the financial security for next year’s conference planned out. We’re excited to see you all again in 2026! The PyOhio Community is incredibly fantastic and we can’t wait to see what you work on this next year!

    • PyOhio Attendees at Meta’s Pyrefly booth, talking with sponsors.
    • PyOhio Attendees at Meta’s Pyrefly booth, talking with sponsors.
    • A view of a table top with Pyrefly stickers scattered around, and printouts for: Meta Careers with a QR code, a CPython free threading logo with a QR Code, and Meta Pyrefly with "Learn More at https://pyrefly.org/pyohio2025"
  • 🔎 Meta Open Source 101🔎 Discover Glean, Meta's open-source system for collecting and querying facts about source code. It enhances developer tools by providing precise answers to code-related questions across multiple languages. 🛠️ Getting Started: To use Glean, clone the repository and build it using the provided scripts. Glean supports various languages, and you can start indexing your codebase to enable powerful code navigation and analysis features. 🔑 Key Features: Glean offers a robust query language, Angle, allowing you to perform complex queries on your codebase. It supports code navigation, search, and static analysis, making it an essential tool for understanding large codebases. 📈 Integration & Usage: Glean integrates with popular development environments and tools, providing seamless access to code insights. Use it to power code search, refactoring, and more. Explore its full potential and contribute to its development: https://glean.software/

  • 🔎 Meta Open Source 101🔎 Discover Captum, a comprehensive model interpretability library for PyTorch. It helps you understand and visualize the decisions made by your machine learning models, enhancing transparency and trust. 🛠️ Getting Started: To use Captum, integrate it with your PyTorch models by installing it via pip with `pip install captum`. Once installed, you can begin exploring various interpretability algorithms to gain insights into your model's behavior. 🔍 Key Features: Captum offers a range of interpretability methods, including feature attribution, layer attribution, and neuron attribution. These tools help you dissect model predictions, identify important features, and understand model biases. 📈 Visualization & Analysis: Captum provides visualization tools to help you analyze and present model insights effectively. Whether you're debugging models or explaining them to stakeholders, Captum's visualizations make complex data more accessible. Learn more about Captum here: https://captum.ai/

Affiliated pages

Similar pages