A .zstd file is a Zstandard-compressed stream written with the spelled-out extension. It is the same format as .zst — the reference tool writes .zst, but .zstd appears often enough from scripts, language libraries and build systems that a converter has to handle both.
Private for supported formats — processed in your browser
Convert supported files
Runs on your device
+
Drop ZSTD files here
Batch files can each use a different output. Nothing uploads for local conversions.
Working inputs include camera RAW, browser-local audio/video, PDF, CBZ/CBR comics, office documents, ebooks, markup, 3D models, structured text, images, and archives.
Zstandard shipped 1.0 in 2016 from Yann Collet at Facebook, and the reference implementation settled on .zst as its suffix. The longer .zstd never went away: it reads unambiguously, several libraries default to it, and the format is identical either way. Nothing inside the file distinguishes them, which is why the extension is the only thing a tool can dispatch on — and why both are registered here as separate inputs.
How ZSTD works
Byte-identical to .zst: the same 28 B5 2F FD frame magic and the same frame structure.
Nothing in the stream records which extension it was saved under, so a reader that supports one supports both.
Compression uses LZ77 matching with finite state entropy coding at levels 1 through 22.
The decompressed size is an optional header field, so a reader must be prepared for it to be absent.
Frames may be concatenated; a conforming decoder reads them in sequence as one logical stream.
When to use ZSTD
Opening a Zstandard file produced by a library or script that writes the long extension
Unpacking build output or a cache artefact compressed with zstd
Reading a compressed dataset without installing the reference tool
Rewriting the contents as ZIP for a recipient whose tools do not know Zstandard
Strengths and limitations
Strengths
Identical to .zst, so nothing about the choice of extension costs compatibility with a decoder that supports the format
Fast decompression makes it practical for data read far more often than written
Permissively licensed with implementations in every major language
Limitations
Two spellings for one format means extension-driven tools must know both, and many know only one
Support in older archive utilities is still patchy
As with .zst, dictionary-compressed streams cannot be read without their dictionary
Compatibility
Any Zstandard-capable tool reads the bytes; whether it recognises the .zstd name is a separate question, and some will not. Novus Convert dispatches on the extension and treats .zstd exactly as .zst, decoding in the browser and writing ZIP, TAR or TAR.GZ.
The compressor stores no file metadata; what survives is whatever the TAR inside carries.