Depositing to a repository

A DOI (a permanent, citable link) is what turns a folder of files into shareable, findable research. Repositories like Zenodo and Figshare mint that DOI for you when you upload a dataset and fill in a metadata form. The form is the slow part, since it wants a title, an author with an ORCID, an abstract, a license, keywords, and the funder and award number behind the work. ResearchOS already holds most of that, so the Deposit flow fills the form in for you and hands you a clean bundle to upload.

You always control the publish

This is the most important thing to understand about the Deposit flow. ResearchOS never publishes anything on your behalf and never talks to a repository for you. It builds the bundle and the metadata locally, in your browser, and hands them to you. The final upload and the final Publish button happen on the repository's own site, with your hands on the controls. The repository, not ResearchOS, mints the DOI.

That means nothing leaves your machine until you decide to upload it, and nothing becomes public until you click Publish on the repository. There is no auto-publish, no background sync, and no surprise release of your data.

Opening Deposit

You can deposit a single experiment or a whole project.

• From an experiment: open the experiment's detail popup and click the Deposit to a repository button in its header. The bundle is that one experiment.
• From a project: open the project and choose Deposit to bundle it as one dataset. The first step lets you multi-select which experiments and which notes go in (notes come from your full note list, since notes are not project-scoped). The bundle then holds each item exported on its own, one combined navigable PDF across everything, the raw re-importable bundle for each experiment, and a markdown rendering of each note.

The three steps

The Deposit dialog walks through curation, metadata review, and handoff.

1. Curate. Choose which sections and attachments go in the bundle, and pick the bundle format (a self-contained HTML page, a PDF, or the raw files). This reuses the same export pipeline that powers a normal experiment export, so the bundle looks exactly like the rest of ResearchOS's exports.

2. Review the metadata. ResearchOS prefills a DataCite-shaped metadata record (the standard repositories speak) and shows it for you to check and edit. See the field list below.

3. Hand off. Pick Zenodo, Figshare, or Other repository. ResearchOS downloads one archive (named {name}-deposit.zip) that holds your curated data and a datacite.json metadata file, then opens the repository's own upload page in a new tab. For Zenodo and Figshare that button jumps you straight to their new-upload form. You drag the archive in, copy the metadata fields into the form, and click Publish there to mint the DOI.

What gets prefilled

The metadata step starts from what ResearchOS already knows, so you mostly confirm rather than type.

• Title: the experiment title, or the project name for a project-level deposit.
• Creator with ORCID: your display name, with your ORCID attached as a name identifier when you have one on file.
• Abstract: seeded from your results, then notes, with the editing-history stamps stripped out. Fully editable.
• License: the one field you must set before handoff, since NIH expects shared data to carry a license and ResearchOS has none to fall back on. The two NIH-friendly defaults (CC-BY-4.0 and CC0-1.0) are surfaced first as recommended.
• Keywords: pulled from the experiment's tags.
• Funder and award number: read from the grant linked to the project, including the funder name, the funder ID and ID type, and the award number and title. These populate the DataCite fundingReference so the deposit carries the grant attribution funders expect. A project-level deposit goes further and lists multiple funders, the primary grant linked to the project plus any other grants that purchases in the project were charged to, deduped so the same grant never shows twice.

After the handoff

The deposit archive is a download. It is never written into your on-disk data folder and never becomes a new sidecar, so depositing leaves your project files untouched. Once the repository mints the DOI, copy it back into ResearchOS wherever you cite the dataset, and record it in your Data Management and Sharing Plan as your funder requires.