Build with mzPeak
Everything around mzPeak is open source — a specification, reference reader/writers across seven languages, converters, a conformance validator, and an in-browser viewer. Pick a starting point.
The canonical format — JSON Schemas, controlled-vocabulary rules, and prose.
Read/write libraries across seven languages — Rust (reference, mzpeak_prototyping), Python, R, C#, JavaScript/TypeScript, C++, and Java — all built on Apache Arrow / Parquet.
Open any .mzpeak directly in your browser, streamed over HTTP — no upload, no backend. Inspect spectra and chromatograms for LC-/GC-MS runs, or render ion images and click pixels for MS-imaging (MSI) datasets.
A language-independent, profile-driven conformance validator. Check a file in the browser — paste an HTTPS or s3:// URL, or upload a .mzpeak (up to 1 GB) — or from the CLI: mzpeak-validate file.mzpeak.
On the roadmap: a direct vendor RAW → mzPeak path (e.g. inside ProteoWizard
msconvert) so every vendor format converts in one step, embedding the original acquisition method as provenance.
See the viewer in action on the Try it page.
Building a tool that reads or writes mzPeak? The format is language-independent — start from the specification and validate against mzPeakValidator. Contributions and feedback are welcome via the HUPO-PSI repositories.