Grant writing prompt builder
A grant application is judged against a rubric, not read for pleasure. Reviewers score it section by section against published criteria, often in minutes. This builder turns your project details and the funder’s criteria into a structured LLM prompt that writes to the rubric — leading with impact, mapping every section to a scoring dimension, and flagging where you still need real numbers. You bring the prompt to your own model and refine the draft it produces.
How it works
You select the funder type and grant size, summarize the project, list the evaluation criteria, and name the required sections plus a target word count. The tool assembles a prompt with a clear role (“expert grant writer”), the funder context, an instruction to organize the narrative around the scoring criteria, and explicit guardrails — no fabricated statistics, placeholders for missing figures, and tone matched to the funder. Everything runs locally; nothing is sent anywhere until you paste the prompt into your chosen LLM.
How different funder types weight the sections
The same project narrative pitched to different funder types needs different emphasis:
| Funder type | What they weight most | Tone |
|---|---|---|
| Government / public agency | Measurable impact, value for money, policy alignment | Formal, referenced |
| Private foundation | Mission alignment, long-term sustainability, community benefit | Values-led, narrative |
| Corporate / CSR | Business relevance, brand fit, ROI | Outcomes-first, concise |
| Academic / research council | Novelty, methodological rigor, investigator track record | Technical, cited |
The builder adjusts tone and section weighting when you choose a funder type. Submitting a foundation-style narrative to a government tender — or vice versa — is one of the most common reasons strong projects are passed over by reviewers.
How to use the evaluation criteria field
Copy the scoring rubric directly from the funder’s call for proposals if you have it. Even rough phrases like “impact (40%), feasibility (30%), sustainability (30%)” tell the model how much space to allocate. If the call lists five criteria with point values, the generated prompt will instruct the model to address each criterion and weight its wordcount accordingly.
If you do not have the explicit rubric, name the criteria you expect based on the funder type. You can refine the draft after you get the actual call.
Tips and examples
- Paste the real call text into the criteria field. The closer the prompt mirrors the rubric, the more closely the draft addresses what reviewers score.
- Keep the project summary concrete. “Reduce ER readmissions for diabetic patients in two clinics” beats “improve healthcare outcomes.”
- Name every required section. Funders reject applications that skip a mandatory section; listing them ensures the prompt covers each one.
- Fill placeholders before submitting. The draft marks where numbers go; never submit with the model’s guesses in place of your real data.
- Have a subject-matter expert review the technical claims. The model structures the narrative; the expertise is yours to verify.