Zstandard
facebook.github.io
- Free plan
- Windows
- macOS
- Linux
- API
- Self-hosted
What Zstandard is
Zstandard is a free, lossless compression algorithm for real-time workloads. It aims for zlib-level speed with better compression ratios, and lets users adjust compression speed and ratio in small increments. Decompression speed remains roughly similar across those settings. For small, related data samples, training can produce a data-specific dictionary to improve compression; the same dictionary must be loaded for compression and decompression, and there is no universal dictionary. The reference implementation includes a C library and command-line tool for creating and decoding .zst, .gz, .xz, and .lz4 files. Its format is stable and documented in RFC 8878, and the reference implementation is licensed under BSD or GPLv2. The project documents builds with Make, CMake, Meson, Visual Studio, Buck, and Bazel, plus installation through vcpkg and Conan. Community-provided language bindings and ports are listed for languages including Python, Rust, Java, C#, JavaScript, Go, PHP, and Swift. Zstandard supports Linux, Windows, and macOS. The project README says Google OSS-Fuzz continuously fuzzes it for security issues.
Who it suits
Zstandard suits developers and teams that need lossless compression in software or command-line workflows across Linux, Windows, or macOS. Its adjustable compression settings and library, CLI, and language bindings offer several ways to integrate it.
Good to know, and what to watch
- Free reference library and command-line utility
- Adjustable compression speed and ratio
- Creates and decodes .zst, .gz, .xz, and .lz4
- Stable format documented in RFC 8878
- Build and package integrations are documented
- Dictionary compression requires matching dictionaries at both ends
- No universal dictionary is available
- Archive encryption and repair are not supported
Plans
Every plan and priceZstandard
Free plan
The verdict
Zstandard offers a free compression library and CLI with configurable speed and ratio settings. Dictionary compression can help with small correlated samples, but requires using the matching dictionary when compressing and decompressing.
What the maker publishes
- API
- The API offers simple methods first and introduces more parameters and control for advanced use. source
- Build and package integrations
- The project documents builds with Make, CMake, Meson, Visual Studio, Buck, and Bazel, and installation through vcpkg and Conan. source
- Build integrations
- The repository documents builds or integration through Make, CMake, Meson, vcpkg, Conan, Visual Studio, Buck, and Bazel. source
- Compression controls
- The compression speed and ratio trade-off is configurable in small increments, and decompression speed remains roughly similar across settings. source
- Dictionary compression
- A training mode builds a dictionary from sample files to improve compression of small, related data; a matching dictionary must be loaded for compression and decompression. source
- Format
- The Zstandard format is stable and documented in RFC 8878. source
- Format
- The Zstandard format is stable, documented in RFC 8878, and has multiple independent implementations. source
- Language bindings
- The homepage lists community-provided bindings and ports for languages including Python, Rust, Java, C#, JavaScript, Go, PHP, and Swift. source
- Latest listed release
- The GitHub releases page lists Zstandard v1.5.7 as the latest release. source
- Library and CLI
- The reference implementation includes a C library and a command-line utility that produces and decodes .zst, .gz, .xz, and .lz4 files. source
- Library and CLI
- The reference implementation provides a C library and a command-line utility that can produce and decode .zst, .gz, .xz, and .lz4 files. source
- License
- The reference implementation is dual-licensed under BSD or GPLv2. source
- Platform builds
- The project documents Windows builds and CMake Universal2 builds for Apple Silicon and Intel Macs. source
- Purpose
- Zstandard is a fast, lossless compression algorithm for real-time scenarios, targeting zlib-level speed with better compression ratios. source
- Security reporting
- The project asks reporters to submit security issues through Meta's Bug Bounty program rather than public GitHub issues or pull requests. source
- Security testing
- The project README says Zstandard is continuously fuzzed for security issues by Google's OSS-Fuzz program. source
- Security testing
- The project says it is continuously fuzzed for security issues by Google’s OSS-Fuzz program. source
- Small-data constraint
- Dictionary training works when a family of small samples is correlated, and the project says there is no universal dictionary. source
- Small-data dictionaries
- Training on sample files creates a data-specific dictionary that can improve compression of small, correlated data samples. source
- Speed and ratio controls
- Compression speed and ratio can be adjusted in small increments, while decompression speed remains roughly similar across settings. source
- Use at scale
- The project says Zstandard is deployed within Meta and other large cloud infrastructures to compress data in various formats and use cases. source
- Use at scale
- The project says Zstandard is deployed within Meta and other large cloud infrastructures to compress data in various formats and use cases. source
- Vulnerability notices
- For significant vulnerabilities not yet public, maintainers may notify selected high-exposure users before public disclosure. source
- Vulnerability notifications
- At its discretion, the project may notify high-profile, high-exposure users about significant undisclosed vulnerabilities before public disclosure and a fix. source
Checked against the maker's pages on 8 Oct 2026.