CrateWorks rolls up an RO-Crate API specification and the open-source tools built around it into one governed home — so researchers, archivists, and developers can pick up any tool on the bench and trust it to fit with the others.
Researchers have been writing one-off scripts and bespoke portals to move research data around for as long as research data has existed. The result is brittle: every archive speaks its own dialect, every tool reinvents the same plumbing, and nothing composes.
RO-Crate fixed the data-format half of that problem by giving us a portable way to describe research objects. CrateWorks is the home for the second half: a shared API that any repository can speak, and a growing kit of tools that all speak it.
Pick up a discovery portal, a bulk-export service, an offline viewer, a field-capture appliance, an audio annotator — and trust that they will work with the next archive you point them at, because they all speak the same protocol.
Two specifications, one reference implementation, many consumers. Anything in the bottom row can be swapped out — they all speak the same API.
The data format. A JSON-LD metadata file that describes a research dataset and everything inside it.
The access protocol. A specification for how any repository serves its RO-Crates over HTTP — discovery, retrieval, deposit, access control.
The reference implementation. A library you mount into a Fastify or Express app, with pluggable storage and external authority for access control.
Discovery portals, downloaders, viewers, capture appliances, annotators — all written against the API, not against any particular repository.
Each tool is independent, open-source, and built to compose with the others through an RO-Crate API.
A standard HTTP API for serving RO-Crate archives.
The specification that ties the workbench together. Defines a portable, implementation-neutral way to discover, read, and deposit research data described with RO-Crate metadata — so the same client can talk to any conformant repository.
Reference implementation of an RO-Crate API.
A Fastify/Express library that mounts the RO-Crate API onto your own app, with pluggable storage (OCFL, object storage, plain directories) and access control wired through external authorities like REMS. Built to follow the PILARS protocols for long-term archival repositories.
A configuration-driven discovery portal for RO-Crate archives.
A Vue 3 frontend that browses research metadata served by any Arocapi-backed repository. Search fields, facets, branding, and navigation are all controlled by a single JSON configuration — point it at a different collection and it adapts.
Pick files across a collection, get a zip in your inbox.
A web app that browses RO-Crate collections hierarchically, lets users pick files with quality-tier filtering (archival vs compressed), then bundles the selection in the background and emails a download link when it is ready.
Local-first audio annotation in your browser.
A waveform editor with automatic voice-activity segmentation, speaker assignment, looping playback, and multi-format export (EAF, SRT, TextGrid, CSV). Everything runs in the browser — your audio never leaves your device.
A field catalogue manager on a Raspberry Pi.
A self-contained Raspberry Pi appliance for capturing and managing RO-Crate catalogues in remote locations where mains power and connectivity cannot be assumed. Bundles a React admin interface, an RO-Crate API, and an image builder for the hardware.
Repository →Browse an RO-Crate archive by double-clicking index.html.
A pure-static web app that walks a directory of RO-Crate collections — items, files, inline audio and image playback, full-text search — without a web server. Designed to run from a USB stick, a DVD, or any cold-storage medium.
Repository →CrateWorks operates under a Minimum Viable Governance framework: a small, public set of charter, code of conduct, steering committee, and per-project governance documents that any project joining the workbench agrees to follow.
Read the governance docs →