Purplelink
Not built yet
A review of your NSF or NIH proposal before the panel reads it. Four reviewers read the proposal against the funder's own criteria and return a report of findings, each one quoting the passage it concerns, with a fix checklist.
Grant Review does not exist yet. This page is a way to find out whether enough applicants want it. The report below is a mock-up on an invented proposal, and nothing on this page is a shipped feature.
Your email and these answers are used only to gauge interest and to tell you if Grant Review gets built. See the Privacy Policy.
What it would do
The design follows Paper Review, which uses four reviewers on a manuscript. The roles below are changed for a proposal.
Reads the proposal against the funder's own criteria. For NSF, those are Intellectual Merit and Broader Impacts, and a proposal that does not address both separately is returned without review. For NIH, the three factors for applications due on or after January 25, 2025: Importance of the Research, Rigor and Feasibility, and Expertise and Resources.
Asks whether the design can support the claims. It looks for a stated minimum effect, a variance assumption, a sample that matches the aim, and an analysis plan for each aim.
Reads the way a panelist in an adjacent field would. It names the strongest objection to the central claim and the sentence that would have to change to answer it.
Checks that personnel months, timeline and budget match the work described, and flags aims that need more time or people than the proposal provides.
An excerpt from a mock-up on an invented NSF proposal, "Low-cost soil-moisture sensors for irrigation scheduling on small farms." No real applicant, program or proposal is behind it. Every finding quotes the passage it is about. The report is not a prediction of funding.
Aim 2 has no testable success criterion.
Quoted: "We will deploy sensors on 24 farms and show a significant reduction in water use."
Problem: The aim states no smallest reduction that matters, no estimate of farm-to-farm variance and no analysis plan, so a panelist cannot judge whether 24 farms is enough.
Surfaced by: Methods and statistics, Skeptical panelist (consensus).
Broader Impacts names no outcome and no partner.
Quoted: "Results will be disseminated to growers through outreach."
Problem: Outreach with no named partner, audience size or measure of adoption reads as an intention.
Surfaced by: Criteria check.
The plan
The plan is $19 for a standard review (one pass, four reviewers) and $39 for a deep review (a second pass in which the reviewers read each other's findings, plus a mock-panel summary). Those are intentions, not final prices, and they may change before launch.
Files would be handled the way Paper Review handles them, where the manuscript and the review are deleted 30 minutes after you first open the result, or after 24 hours if you never open it. The details would be stated on the product page before launch, and the Paper Review data-handling section shows the current wording.
Limits
Their criteria are public and specific, so a review can be checked against them. Other agencies and foundations would follow if there is demand.
The three-factor framework applies to research project grant applications due on or after January 25, 2025. The criteria are described on the NIH page for the framework. Some program types still use other criteria, which the review would ask you to confirm.
Your email address and your three answers, kept by Netlify Forms. They are used to decide whether to build Grant Review and to write to you once if it launches.
Purplelink LLC, a one-person studio run by Benjamin Ampel, a researcher who writes and reviews proposals as well as papers.