Loading page
About this tool
Build a reference letter prompt with specific examples, an honest percentile claim from rank and cohort size, and a per-section word budget.
The Recommendation Letter Prompt Builder converts what you personally observed into an AI prompt that produces a specific, defensible reference letter instead of a page of superlatives. It calculates the strongest percentile claim your cohort size honestly supports — rank 3 of 120 is 2.5%, which rounds up to 'top 5%', never down to 'top 1%' — and splits your target length across the six sections of a letter so each part gets a word allowance. Each example you supply is scored for whether it names an artefact, states an outcome and offers a comparison.
Open Recommendation Letter Prompt Builder on AltFTool — it loads instantly in your browser.
Paste or type the text you're working with.
Choose how it should be transformed or analyzed.
Copy the finished text into your document or post.
The claim is rounded up to the next conventional band, so the letter never overstates the rank.
Under 20 people the tool drops percentages entirely and writes out the real numbers.
Each example is checked for a named artefact, a stated outcome and a peer comparison.
One page for an employment reference, one to two pages for academic admissions. In words that is roughly 300 to 500 for a job reference and 500 to 800 for a graduate application; at 12pt single-spaced with one-inch margins a page holds about 500 words. This tool enforces the range for the letter type you pick.
Only if the cohort is large enough for that to be arithmetically possible. In a class of 40 the best possible rank is 1, which is 2.5% — so 'top 5%' is the strongest honest claim. This tool computes the exact percentage from rank and cohort size and rounds up to the next conventional band, and it refuses to produce a percentage at all for cohorts under 20.
Specific incidents the writer personally witnessed, with a named course, project or artefact and a stated outcome. Committees discount adjectives like 'brilliant' and 'exceptional' that arrive without evidence. Stating plainly what you did not see also builds trust — a short honest letter reads better than one that overclaims familiarity.
Research on academic and medical reference letters has repeatedly found that letters for women contain more communal terms — warm, helpful, nurturing, hard-working — while letters for men contain more agentic terms such as ambitious, independent and brilliant, even for candidates with comparable records. The prompt this tool builds explicitly instructs the model to describe the same evidence in neutral terms regardless of the candidate.