Research grant search that answers a plain request with five research funding opportunities, each eligibility, cost share and limit cited to the sponsor's own solicitation
Five high-fit calls in minutes, each limit and cost share cited to the sponsor's own text.
See one case, screen by screen ↓Research grant search is finding the open funding calls a faculty member can actually apply to: calls that fit the work, and whose eligibility, cost sharing and limit per institution allow a proposal. Research Grant Search answers a researcher's plain request with five calls, each with those three rules quoted from the sponsor's full solicitation and a line on why it fits in her own words. People decide whether to apply, and the research office runs any internal competition.
AI in education: a buyer's guide to agentic AI for colleges, universities and school districts →
Updated
Research grant search is the work of finding the open calls a faculty member can actually apply to: the topic has to fit, and so do three rules that sit in the sponsor's solicitation — who may apply, whether cost sharing is required, and how many proposals one university may send. A grant search engine or funding opportunity database matches on keywords and shows a summary its vendor wrote. The funding opportunity announcement itself runs 20 to 130 pages, and on NSF calls the limit sits in section IV, "Limit on Number of Proposals per Organization".
So a research development officer or the researcher opens each solicitation by hand, checks the limited-submission calendar, counts interest by email and, when a call is capped, runs an internal competition about 90 days before the sponsor's deadline. A cap noticed late means faculty write proposals that cannot all go; a cost share noticed late reaches the department research administrator the week before routing. Research Grant Search puts seven agents in front of that reading, and leaves every decision to a named person.
Estimated split, in hours, of the work for one researcher's search and one limited-submission call, by hand and with the agents — our arithmetic, kept modest.
Seven agents read her approved profile and every page of each solicitation, tick eligibility, cost share and limits with the clause, and open the way to internal competitions; people decide.
Dr. Rafaela Quist, associate professor of civil and environmental engineering at Halden State University, needs her next bridge monitoring grant. Here is Wednesday, October 21, 2026, screen by screen, in the working solution.
The Profile agent drafted her profile from 23 papers, 4 past awards and her September 2026 biosketch, and nothing was used for matching until she approved each of the 7 lines. Two student co-authors were found and left out; their names are never stored.
Used only for matching and alerts · private to her and the research office.
She types "funding for bridge corrosion monitoring, under $500K, deadlines after March". The request is shown back as "Understood as" chips she can edit and rerun, with no filters panel. The Matcher narrows 45,012 open calls to 318 on the topic and 41 within award and deadline, and five come back ranked by fit, each with three ticks: Eligible, Cost share and Limited submission.
5 calls fit from 45,012 open · 5 solicitations read in full · ranked by fit to Dr. Quist's own words.
The FHWA call's cost-share tick reads "20% required: $90,000 on a $450,000 request", cited to §III.B. On NSF 26-547, the limited-submission tick opens section IV of the 22-page solicitation with the clause marked: an organization may take part in no more than two proposals, and a subaward of 20 percent or less does not count. The agent read it with confidence 0.97, above the 0.90 threshold.
She asks whether a subaward counts against the limit, then which calls need cost share. The answers come only from the five solicitations, each with its clause: FHWA needs $90,000, the state DOT call a $60,000 match with in-kind allowed, and NSF 26-547 does not allow voluntary cost share.
A question the agents cannot answer goes to the research development office, not to a guess.
The database summary for NSF 26-547 says "Limit on proposals: none specified"; section IV says 2 per institution, and the agents used the solicitation. The call is on Halden's limited-submission calendar as LS-2027-031, internal pre-proposals due January 9, 2027. She clicks I'm interested, and five faculty now want the two slots; on the FHWA call she says she will apply.
The database summary is wrong here. The agents used the solicitation.
Theo opens LS-2027-031, which sits on the calendar with 5 interested over a limit of 2. The agents create the 3-page pre-proposal form, draft the announcement with section IV quoted, set reminders for December 28 and January 6, and hold the panel window January 11–20, every date planned back from the April 8, 2027 sponsor deadline.
Before the announcement reaches 12 faculty — 5 interested and 7 high-fit researchers who allow suggestions — the Director checks that the limit matches the solicitation, the dates run back from the deadline and every recipient allows research-office contact. She sends it; Theo, who opened it, cannot.
On a separate call, LS-2026-118, the Ardent Foundation allows one nomination. Six pre-proposals were read against the sponsor's three criteria with page citations; one reviewer was not seated for a conflict and one recused. Theo recommended Dr. Malik Sorensen's $612,000 shared X-ray diffractometer, panel score 4.6 of 5, and Dr. Voss approves.
Dr. Quist's intent, IN-27-0412, arrives with the three cited ticks carried over and routing to pre-award due April 8, 2027. Priya starts proposal record PR-27-0412 in the research administration system and sends the $90,000 cost share to the department chair, then sponsored programs.
The request, the interest, the intent, the competition opened, the announcement checked and sent, the nomination approved, the proposal record and the cost-share request are on the trail with who, when, why and the evidence, people and agents alike. The audit pack exports with pre-proposal text and student names left out.
The same morning, seen by the five people who carry it — what their day looked like, and what it looks like now.
Seven agents read her approved profile and every page of each solicitation, tick eligibility, cost share and limits with the clause, and open the way to internal competitions; people decide.
Drafts each researcher's profile lines from papers, past awards and the biosketch, each with the source it came from.
Brings in new and amended calls every morning from federal portals, foundation databases, sponsor alerts and state programs, and keeps the full solicitation, not only the summary.
Filters by award and deadline, ranks by fit to the researcher's own words, and writes the why-it-fits line that quotes her sentence beside the call's.
Reads the whole solicitation and answers three questions — eligible, cost share, limit per institution — quoting the clause and its section, and flags where the database summary disagrees.
Finds a capped call on the limited-submission calendar, counts interest and slots already used, and plans every internal date back from the sponsor's deadline.
Checks reviewers against applicants for co-authorship in the last 4 years, shared awards and the same department, and reads each pre-proposal against the sponsor's criteria with page citations.
Sends each consenting researcher a weekly digest of new calls above her fit bar, with the ticks and the fit reason cited, and builds targeted-alert lists for the office.
Approves every selection; the Director checks every notice. The agents propose; a named person decides.
Ask in plain words. Answers come from the sponsor's own text, with the clause, and the right screen opens.
Does a subaward count against the limit?
Yes, unless it is small. A subaward of 20 percent or less of the total budget does not count against NSF 26-547's limit. A larger one uses one of Halden's 2 slots.
Which of my five calls need cost share?
Two of the five. The FHWA call needs 20% of the federal request, $90,000 on $450,000. The state DOT call needs a 20% match, $60,000, and in-kind counts. NSF 26-547 does not allow voluntary cost share at all.
Who else wants to apply to NSF 26-547?
Four faculty have shown interest, from Civil and Environmental Engineering, Mechanical Engineering and Materials. Their names are visible only to the research office. Halden can send 2 proposals, so Theo Marchetti runs an internal competition; pre-proposals are due January 9.
Which limited-submission deadlines need action this week?
Three. LS-2027-031, NSF 26-547: interest is over the limit, so the competition should open by October 23. LS-2026-118, Ardent Foundation: the selection is waiting for Dr. Helena Voss, due October 23. LS-2027-011, Tessaly Foundation: its approval is 2 days late and escalated to the Director.
13 screens from the working solution, on its sample data. Pick one to see it large.
The thresholds, the internal deadlines and who approves what are settings. A systems administrator proposes a change; the Director approves it, and it goes on the record.
| Setting | Default | Choose from |
|---|---|---|
| Eligibility confidence threshold | 0.90 | 0.80–0.98; below it a tick goes to the research development office |
| Internal deadline | About 90 days before the sponsor deadline | Proposed by the administrator, approved by the Director |
| Pre-proposal | 3 pages and a 2-page CV | Set per competition |
| Subawards | Follow the solicitation | If it is silent, count any subaward |
| Review panel | 3 reviewers | Not seated if a co-author in the last 4 years, a shared award or the same department |
| Open a competition after interest passes the limit | 2 business days | Escalate to the Director when late |
| Approve a selection | 3 business days | Escalate to the Director when late |
| Override a tick | Research development officer, with a reason code | OV-1 to OV-4; reviewed weekly by the Director |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
Sources: NSF National Center for Science and Engineering Statistics, Higher Education R&D Survey, fiscal year 2024 ($117.7 billion; $64 billion federal) · industry data: limited submissions are 16–20% of US government funding opportunities · Federal Demonstration Partnership faculty workload survey (42% of research time on administration) · NSF solicitations set limits per organization in section IV · 2 CFR 200.306 (cost sharing) and 200.334 (records) · FERPA, 34 CFR Part 99 · "target" = design goal, measured in the live solution · "demo" = seen in the working solution, on its sample data · "typical" = published figures · "estimated" = our arithmetic. Halden State University, the people named on this page, the foundations, and the calls, solicitation numbers and clauses shown are fictional — characters and sample data in the working solution.
Research grant search is finding the open funding calls a faculty member can actually apply to: calls that fit the work, and whose eligibility, cost sharing and limit per institution allow a proposal. Research Grant Search answers a researcher's plain request with five calls, each with those three rules quoted from the sponsor's full solicitation and a line on why it fits in her own words. People decide whether to apply, and the research office runs any internal competition.
A grant search engine or grant search database matches on keywords and shows a summary. Here the agents read every page of each solicitation and quote the clause behind each tick, and when the database summary disagrees — NSF 26-547's summary says no limit, section IV says 2 per institution — they use the solicitation and say so.
The Eligibility agent reads the whole funding opportunity announcement and answers three questions: is the researcher eligible, is cost sharing required and how much, and is there a limit per institution. Each answer quotes the clause and its section or page. Below 0.90 confidence the tick is not shown, and the research office reads the solicitation.
A limited submission is a call that caps how many proposals one institution may send. When a call on the shortlist is capped, the Internal Rules agent finds it on the limited-submission calendar, counts interest against slots and plans every internal date back from the sponsor's deadline. The research development officer opens the competition, the Director checks the announcement before it is sent, a conflict-checked panel reviews, and the Vice President for Research approves the selection.
No. The agents read, rank, cite and plan; they never apply, open a competition, send a notice, change a profile or select. The researcher decides whether to apply, the Director checks every notice, and the Vice President for Research approves every selection, with the maker never the checker.
A profile is used only after the researcher approves each line, and she can switch consent off or withdraw it, which pauses matching and deletes the profile after 30 days. Faculty see only a count of who is interested in a capped call. Student co-authors' names, grades and enrollment are never stored.
Yes. The confidence threshold, the internal deadline, the pre-proposal length, how subawards count, panel conflict rules, the clocks and escalations, and retention are settings. A systems administrator proposes a change, the Director approves it, and it is on the audit trail.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your limited-submission policy, your calendar and a sample of this cycle's solicitations — and it goes live once every quality gate has passed. We will walk you through it on your own calls first.
Every proposal reviewed inside the five-day window, every value tied to its clause.
1-min demoSubrecipient Risk Assessment & MonitoringRisk scored before signing and every invoice line checked before the university pays.
1-min demoIRB Protocol Pre-ReviewEvery protocol back to its investigator on the first day, each fix tied to its rule.
1-min demoWe'll run Research Grant Search on this cycle's solicitations and your limited-submission calendar.