For GRFP applicants drafting the research plan
NSF GRFP Research Proposal Writer
A two-page draft of your NSF GRFP research proposal, built from your own idea or paper, with a testable hypothesis, aims and fallbacks.
Free to start. No card needed.
How to write the NSF GRFP research proposal yourself
Start with one question you could make real progress on in a few years, and check it against NSF's ineligible research rules first. NSF's additional FAQ1 says it screened the research plan, not the personal statement, to judge eligibility in the 2026 competition (applications in fall 2025). Then turn the question into a hypothesis that could prove wrong, and state it outright. NSF's tips for applicants2 say "Use hypotheses, figures, and references." On one 2025 ratings sheet an applicant shared, a reviewer wrote that an explicit hypothesis "would have further strengthened the proposal."
Give each aim a short block: the approach, the result you expect, and what you'll switch to if the main method fails. Build in a comparison, such as a control, that rules out the obvious other explanation. NSF's research plan template3 says reviewers look for your ability to "plan and conduct research" and "solve problems," and a fallback shows both. List the "Resources needed" too: the lab, data or training the work depends on. Reviewers will generally know your field but probably not your topic, NSF's tips note, so explain each method plainly.
NSF's rules for the 2026–27 cycle4 require separate sections headed exactly "Intellectual Merit" and "Broader Impacts", with a line break between them. Under the first, say what your field gains from the answer and why you're ready to do this work. Under the second, name who could benefit from this project and how you'll share the results. References and figures count toward the two pages. Cite journals by abbreviated title, with no URLs or DOIs. Give every figure alt text, then check the PDF with the format checker.
From NSF's template
What reviewers look for in a GRFP research plan
NSF's research plan template3 lists what reviewers look for in this statement, quoted below word for word, and adds: "You do not need to address every bullet." Here's how to show each one. The example lines are illustrative, written for the sample drought plan shown further down.
| What reviewers look for, in NSF's words | How to show it | Example line (illustrative) |
|---|---|---|
| How your research advances knowledge and understanding in your field | Say what your field learns whether the hypothesis holds or fails | "Either result shows whether drought tolerance can travel with soil fungi." |
| Whether the rationale for your research plan is clear and compelling | Open with the problem, what is known and the gap, before any methods | "Root fungi can change how plants take up water [1], but whether dry-site fungi pass on drought tolerance is unknown." |
| Your ability to: plan and conduct research, work independently and/or as part of a team, solve problems | Name concrete methods, point to work you've done yourself, and give each aim a fallback | "If regrowth doesn't differ, I will compare survival instead." |
| Your familiarity with relevant literature | Cite the few studies that define the gap, and say where your question goes further | "Work on crop plants [2, 3] leaves open whether wild grasses respond the same way." |
| Whether your research plan is logical | Make each aim test something the hypothesis predicts, so the aims build on each other | "Aim 1 tests whether dry-site fungi speed regrowth. Aim 2 tests how they do it." |
Quoted from the research plan template NSF posted for the 2026–27 cycle (NSF 26-526), checked October 29, 2026.
How it works
How does River draft your GRFP research proposal?
The NSF GRFP research proposal, officially the Graduate Research Plan Statement, is a two-page plan for the research you'd pursue in graduate school. Start with whatever you have: an idea, your lab's project, an abstract, or a paper, poster or old draft dropped into the box. River reads it all, then drafts a specific title, the gap, an explicit hypothesis, two or three aims with fallbacks, the resources you need and both required sections. It builds from your own words and keeps your strongest phrases.
Before drafting, River checks your idea against NSF's rules on ineligible research, such as goals "directly human disease- or health-related." If a rule may apply, your doc opens with it quoted. River also names a Major Field of Study and two subfields from NSF's list. That choice matters: under NSF 26-5264, an application filed under a field that doesn't fit the plan is returned without review. River never invents data, citations, collaborators or equipment: gaps become [bracketed slots], and each sentence it wrote carries a comment asking you to rewrite it.
The 2026–27 cycle has closed. Applications were due October 19 to 23, 2026, and NSF expects to announce results by early April 2027. River's draft follows NSF 26-526, so if you're applying next year, recheck the rules when NSF posts its next solicitation (this one appeared August 25, 2026). Unsure about your topic? Start with the eligibility checker. Then draft your other statement with the personal statement writer so both tell one story, and run the mock review before you submit.
How it works
From an idea to a two-page plan
Paste or drop
Paste an idea, your lab's project, an abstract or your interests, or drop a paper or draft.
River checks eligibility
It reads everything, checks your topic against NSF's research exclusions, and settles on one testable question.
Get your plan
About two pages land in your doc, with bracketed slots for missing facts and a field suggestion.
Make it yours
Rewrite every marked line in your voice, fill each slot, then ask for a reviewer's read.
Do root fungi from dry grasslands help wild grasses survive drought?
Example plan · Life Sciences, Ecology · about 1,050 words
- HHypothesisGrasses grown with fungi from a dry site regrow faster after drought than grasses grown with fungi from a wet site.
- 1Aim 1 · Greenhouse drought testExpected: faster regrowth. If not: compare survival instead.
- 2Aim 2 · How the fungi change water uptakeNeeds: [root imaging equipment you can use]
- IMIntellectual MeritTests whether drought tolerance can travel with soil fungi
- BIBroader ImpactsGrows from your school garden club · [partner school]
What you'll get
A two-page research plan drafted from your idea or paper
- A two-page plan draft: title, gap, hypothesis, and two or three aims with fallbacks
- Intellectual Merit and Broader Impacts under NSF's exact headings, tied to your project
- Your best-fit field and two subfields, a figure idea with alt text, and what's still open
Before and after
Your notes in. A two-page plan out.
Senior in ecology. In my lab I sampled root fungi at a dry grassland and a wet one, and I ran DNA extractions on 60 samples. I want to know if fungi from the dry site help plants survive drought. I also volunteer with a school garden club.
Do root fungi from dry grasslands help wild grasses survive drought? Background and gap. Longer droughts threaten many grasslands [1]. Root fungi can change how plants take up water [2], but whether fungi from dry sites carry drought tolerance to new plants is [unclear: check the literature]. Hypothesis. I hypothesize that grasses grown with root fungi from a dry site will regrow faster after drought than grasses grown with fungi from a wet site. Aim 1. Building on my 60 DNA extractions from both sites, I will grow [number] seedlings with each fungal community and withhold water for [length of drought].
Each sentence River wrote, like the background line, carries a comment asking you to rewrite it. Citations [1] and [2] point to numbered slots under References until you add real sources.
What lands in your doc
What's in your research plan draft
- A specific title, the background and gap, one research question and an explicit hypothesis
- Two or three aims, each with its approach, expected results and a fallback if it fails
- In each aim, a comparison that rules out the most obvious other explanation for your result
- Resources needed, such as the lab, equipment, data and training, as you gave them or as slots
- Intellectual Merit and Broader Impacts under NSF's exact headings, both tied to your own project
- A check against NSF's ineligible research rules, with the rule quoted at the top if it applies
- A closing list: your Major Field, two subfields, a figure idea with alt text and every open slot
Questions GRFP applicants ask
Common questions
What goes in the NSF GRFP research proposal?
NSF's research plan template3 asks for an original research topic with a title, the research idea and proposed approach, and the resources needed. It also requires separate sections headed exactly "Intellectual Merit" and "Broader Impacts". Other headings are allowed, and sections can come in any order. Within two pages, a strong plan adds the background and gap, an explicit hypothesis, and two or three aims with expected results and fallbacks.
Do I have to do the research I propose?
No. NSF's applicant FAQ5 says "GRFP funds people, not specific projects. Your plan demonstrates potential, not a binding commitment." The field does count: under NSF 26-5264, accepting the fellowship commits you to a graduate program "consistent with the field of study" in your application. Major changes in scope later in your graduate career need NSF approval.
Is my research topic eligible for the GRFP?
Only if it avoids the exclusions in NSF 26-5264. Research whose goals are "directly human disease- or health-related" is out. So are applied studies of crop production or food safety, research meant to influence specific policies, and research on conservation practice or wildlife management. Bioengineering that mainly advances engineering knowledge can qualify, but only under Engineering. River quotes any rule that may apply, and NSF makes the final call.
Can I use AI to write my GRFP research plan?
Nothing in NSF 26-5264 bans it, but you must certify that both statements are your "own original work." The solicitation also quotes NSF's policy that "The responsibility for proper scholarship and attribution rests with the authors of a proposal." So treat any AI draft as a starting point. River works only from your material and marks every sentence it wrote for you to rewrite in your own words.
How do I write my GRFP research proposal with River?
Paste your idea, your lab's project or an abstract into the box at the top of this page, or drop a paper or old draft. If you only have interests, River picks the most feasible question and offers two other directions. It checks eligibility, then drafts about two pages into your free River workspace, where you rewrite each marked line and fill every slot.
Will River make up references or preliminary data?
No. River never invents data, results, citations, collaborators, lab names, equipment or numbers. NSF's applicant FAQ5 allows preliminary results "if they help demonstrate your preparation and potential," so River uses yours exactly as you state them and says whose they are. It formats the references you give with abbreviated journal titles and no URLs or DOIs, and leaves a numbered slot where a source is missing.
Is the NSF GRFP Research Proposal Writer free?
Yes, it is free to start. Your research plan opens in a River workspace, where you can edit it, share it or ask River to change it. Paid plans add more AI use and access to the most advanced models. River's privacy policy says it does not use your documents or drafts to train AI models. Still, ask your advisor before uploading a lab's unpublished results.
Your next step