How to Use DMPTool for NSF/NIH DMPs

DMPTool is the free, web-based data management plan authoring service run by the California Digital Library (CDL) through its University of California Curation Center (UC3), used by researchers across the United States to write funder-compliant data management plans (DMPs). This guide walks through the actual DMPTool workflow, step by step, for a researcher or research administrator building an NSF or NIH data management plan — from account setup through exporting the finished plan — and explains where DMPTool’s funder templates fit against the current NSF and NIH format requirements.

Before you start: what NSF and NIH actually require

DMPTool’s job is to walk you through a funder’s specific template; it does not change what the funder requires. As of 2026, both agencies’ formats have moved away from older free-text conventions:

NSF. Effective for proposals submitted on or after April 27, 2026, NSF moved data management plans from an uploaded PDF supplementary document into a structured webform inside Research.gov, tailored to the proposal’s primary Directorate (proposals submitted before that date could still use the older PDF-upload format). This replaced the long-standing convention of a free-standing two-page PDF narrative. Every NSF directorate still requires a Data Management Plan for proposals that generate data, but the current mechanism for producing it is the structured Research.gov form, not a document you draft and upload independently.
NIH. Effective May 25, 2026, NIH’s Data Management and Sharing (DMS) Plan moved from a two-page narrative to an abbreviated “2026 Pilot” format: Yes/No questions about data-sharing practices, a maximum 300-word explanation for any limited sharing, and a short (maximum 100-word) table of anticipated data types and repositories. This sits under NIH’s 2023 Data Management and Sharing Policy, which requires a DMS Plan for essentially all NIH-funded research that will generate scientific data, submitted at the application stage — a separate obligation from NIH’s Public Access Policy, which governs peer-reviewed manuscripts rather than underlying data. See CASRAI’s NIH vs. NSF Data Management Plans guide for a full comparison of how the two agencies’ requirements actually differ.

DMPTool’s own template library tracks these changes on the DMPTool side (its NSF and NIH templates reflect the current formats, including retirement of older institute-specific templates that no longer match NIH’s current structure), but neither DMPTool nor any third-party tool submits the plan into Research.gov or NIH’s ASSIST/eRA Commons system for you. DMPTool is where you draft, structure, and get institutional sign-off on the plan’s content; the funder’s own portal is where the final version is ultimately entered or attached at submission. Treat DMPTool as your drafting and review workspace, not your submission portal.

Step 1: Create a DMPTool account

DMPTool does not require an institutional login. If your institution participates in DMPTool (most US research universities do), you can typically sign in via institutional single sign-on (Shibboleth); if your institution isn’t a registered participant, or you have no institutional affiliation, you can create a standalone account directly at dmptool.org. A standalone account has full access to create and write plans — institutional participation mainly affects whether you see your institution’s customized guidance text and locally branded templates inside a plan.

Step 2: Start a new plan and set the project title

From the dashboard, select “Create plan.” DMPTool’s own quick-start guidance is explicit here: enter a title for your research project, and if you’re applying for funding, use the same project title that appears in your actual proposal — this keeps the plan’s identifying metadata consistent with what the funder sees elsewhere in your application.

Step 3: Select your organization

Choose your research organization from DMPTool’s lookup list. If your account is already associated with a participating institution, this field is pre-populated automatically; you can clear it and search again if it’s wrong, or select “No organization is associated with this plan” if none applies. Selecting the correct organization is also what surfaces your institution’s own guidance text and any locally customized templates, where your library or research office has configured them. DMPTool matches organizations against the Research Organization Registry (ROR) and the Crossref Funder Registry so the identifier, not just the display name, travels with your plan — see CASRAI’s guide on DMPTool’s ROR and Funder Registry integration for exactly what that auto-fill does and doesn’t cover.

Step 4: Select NSF or NIH as your funder

Select your funder from DMPTool’s supported list. For NSF and NIH specifically, choosing the agency can surface a secondary dropdown, since DMPTool cites NSF (and DOE) as examples of funders whose individual directorates or programs carry different requirements — picking the correct sub-program is what determines which specific question set and guidance text you get in the next step. If neither NSF nor NIH applies to your project, checking “No funder associated with this plan” gives you a generic template instead. There’s also an optional checkbox to mark the plan as a “mock project for test, practice, or educational purposes” — useful for training or a class exercise, but note (see Step 7) that a plan marked this way cannot be registered with a persistent identifier.

Step 5: Write the plan

The plan editor is organized into tabs:

Project Details — administrative information about the project, plus space to select organizational guidance (DMPTool supports up to six guidance selections from participating institutions).
Collaborators — list your PIs and data managers, and invite specific people by email with read, edit, or admin permissions on the plan. This is where you add co-investigators or a data steward as actual contributors rather than just naming them in the text; each contributor can also record their ORCID iD and affiliation if it differs from your own institution.
Write Plan — the core authoring step: answer the funder- or institution-specific questions section by section, with guidance panels alongside each question explaining what NSF or NIH is actually asking for. This is where NSF’s current Directorate-tailored question set, or NIH’s 2026 Pilot Yes/No-plus-short-answer format, is presented to you inside DMPTool’s editor.
Research Outputs — a section for documenting the specific datasets, code, or other outputs your project expects to generate.
Download — where you export the finished plan (see Step 6).

If your library or research office has enabled the service, a “Request feedback” option is available from within the plan: it routes your draft to institutional support staff, who can leave comments next to specific questions, with an email notification once their review is complete. For a look at what a completed plan actually looks like section by section, see CASRAI’s Data Management Plan Worked Examples.

Step 6: Export or download the plan

DMPTool’s Download tab lets you export your plan in multiple formats, with customizable PDF formatting options for the exported document. For an NSF or NIH submission built around the funder’s own structured webform, you’ll typically use the DMPTool draft to work out and refine the plan’s actual content with your co-authors and (if available) institutional reviewers, then transfer or attach the finished text into NSF’s Research.gov webform or NIH’s application system as required — DMPTool does not submit the plan into either agency’s system directly.

Step 7: Register the plan for a DMP ID (optional, but recommended)

DMPTool can register a completed plan and assign it a persistent identifier called a DMP ID — a DOI, issued through DataCite, that uses the “OutputsManagementPlan” resource type so the plan itself becomes a discoverable, citable object (for example, in DataCite Commons) rather than a document that only exists inside your institution’s files. To register a plan and generate a DMP ID, DMPTool requires that you:

have answered at least 50% of the plan’s questions,
have identified a funder on the Project Details tab,
have linked your DMPTool account to your ORCID iD (via the account’s Third-party applications page), and
not have the plan marked as a mock/test project.

A registered DMP ID is what allows a plan to be linked programmatically to its eventual outputs (datasets, publications) as part of the broader push toward machine-actionable data management plans built on the RDA DMP Common Standard. It’s a genuinely useful step for a funded project, but it is separate from, and not a substitute for, actually submitting your DMP content through NSF’s or NIH’s own systems.

Plan visibility: who can see your DMP

New plans default to Private, visible only to you and anyone you’ve added as a collaborator. Public and Organizational visibility settings are both available, but DMPTool requires at least 50% of the plan’s questions to be answered before you can switch to either of those broader visibility levels — a safeguard against sharing a near-empty draft as though it were a finished plan.

DMPTool for NSF vs. DMPTool for NIH: what actually differs inside the tool

The account-creation, plan-creation, writing, collaborator, and export steps above are identical regardless of which agency you’re writing for — the difference is entirely in which template and question set you select in Steps 4 and 5. In practice:

NSF plans in DMPTool present the funder’s Directorate-specific structured question set (reflecting NSF’s April 2026 move to a Research.gov webform), and may prompt you to select a specific program or directorate sub-template.
NIH plans in DMPTool present the agency’s current DMS Plan question set (reflecting NIH’s May 2026 “2026 Pilot” abbreviated format: Yes/No questions, a short limited-sharing explanation, and a short data-types/repositories table), rather than the older two-page narrative format.

If your project has funding from both agencies, or from NSF/NIH plus another funder, DMPTool lets you create a separate plan for each, or duplicate a plan and adjust it — there’s no requirement to force one project into a single template if your actual funding situation doesn’t fit that.

Frequently asked questions

Do I have to use DMPTool for an NSF or NIH data management plan?

No. DMPTool is a free authoring aid, not a mandatory NSF or NIH system. Some institutions strongly encourage or standardize on it because it keeps templates current and enables institutional review, but you can write a compliant DMP without it, provided you meet the funder’s actual current format (NSF’s Research.gov webform, NIH’s DMS Plan format).

Does DMPTool submit my plan to NSF or NIH directly?

No. DMPTool is where you draft, structure, and review a plan’s content. The finished content still has to be entered into NSF’s Research.gov data management plan webform, or attached/entered through NIH’s application system, as part of your actual proposal submission.

Do I need my institution to participate in DMPTool to use it?

No. You can create a standalone DMPTool account even if your institution doesn’t participate. What you lose without institutional participation is access to your institution’s own customized guidance text and branded templates inside the plan editor — the core funder templates and writing workflow are the same either way.

What is a DMP ID, and do I need one?

A DMP ID is a DOI, issued via DataCite, that DMPTool can assign to a completed, registered plan so the plan itself is a discoverable, citable, machine-resolvable object. It isn’t required to write or submit an NSF or NIH DMP, but it’s a low-effort step (requiring 50% of questions answered, a funder identified, your DMPTool account linked to ORCID, and the plan not marked as a test) that supports machine-actionable data management practices going forward.

How is DMPTool different from DMPonline?

Both run on the same open-source DMPRoadmap platform, but they’re operated by different organizations for different funder ecosystems: DMPTool (CDL/UC3) is curated around US funders including NSF and NIH, while DMPonline (the UK’s Digital Curation Centre) is curated around UKRI, Wellcome Trust, and Horizon Europe. See CASRAI’s full DMPTool vs DMPonline comparison for the complete breakdown.

Can I edit a plan after I’ve registered it for a DMP ID?

DMPTool’s own registration requirements are about the state a plan needs to be in before you can register it (50% of questions answered, funder identified, ORCID linked, not a test project) rather than freezing the plan permanently at that point, but always check DMPTool’s current documentation or your institution’s DMPTool support contact before relying on edit behavior for a plan you’ve already registered.

References

DMPTool, Quick Start Guide for Researchers (dmptool.org/quick_start_guide).
DMPTool Blog (blog.dmptool.org).
DataCite, “What is a DMP ID?” and “A Brave New PID: DMP-IDs” (support.datacite.org/docs/what-is-a-dmp-id; datacite.org/blog/announcing-dmp-ids).
NSF, Proposal & Award Policies & Procedures Guide (nsf.gov/policies/pappg) and NSF data management plan requirements.
NIH, Data Management and Sharing Policy and 2026 Pilot DMS Plan format guidance (grants.nih.gov; sharing.nih.gov).