Pro · Expert Letter Studio
In the expert's voice —
provably independent.
Every candidate letter writer is checked for real independence — co-authors and colleagues excluded — then a structured skeleton with genuine placeholder slots for their own voice, cross-checked against a similarity matrix that catches templated praise before it's filed.
Start your trialCo-authors and colleagues excluded automatically.
What each draft grounds.
Provably independent, not just plausible.
Every candidate is checked against your file's own co-authors and colleagues — a shared publication or employer excludes them automatically, so the ones that remain are genuinely arm's-length.
A skeleton, not a draft to sign unedited.
Real placeholder slots for the recommender's own specific claims and comparative context — structured to be filled in their voice, not generic praise ready to be rubber-stamped.
Catches templated language before it's filed.
A cross-petition similarity check flags reused phrasing — the same templated-praise pattern Merito Audit already catches — so a letter never reads as copy-pasted.
A flattering letter isn't the same as a credible one.
A co-author's praise reads as biased, not credible.
A glowing letter from someone with a shared publication or employer relationship gets weighted differently than genuinely independent testimony.
"A truly exceptional talent" says nothing specific.
Generic superlatives don't give an officer anything concrete to weigh — a letter needs named, verifiable claims.
Reused phrasing across petitions is a real red flag.
A phrase that shows up in multiple unrelated petitions this quarter reads as templated, undermining the letter's credibility.
Why a chatbot can't verify a letter writer's independence.
A guess and a verified-independent, evidence-grounded letter are not the same kind of answer.
It can't check for co-authorship or shared employment.
Verifying genuine independence requires cross-referencing real publication and employment records — not something a single conversation can do.
Left alone, it drifts toward generic superlatives.
Ask a chatbot to draft a reference letter and it tends toward flattering, generic language — exactly the pattern that reads as templated.
It has no visibility into other petitions.
Catching reused phrasing across a cross-petition corpus isn't possible for a tool with no access to that corpus.
Independent, specific, and provably not templated.
How a letter gets drafted.
Candidates get checked for independence.
Co-authors and same-employer connections are excluded automatically, leaving genuinely arm's-length candidates.
A skeleton drafts with real placeholder slots.
Structured for the recommender's own specific claims and comparative context, not generic pre-written praise.
A similarity check runs before filing.
Cross-petition comparison flags any reused phrasing before it becomes a credibility problem.
The recommender reviews and signs in their own voice.
The letter is meant for their edit and signature, never for verbatim, unedited use.
DRAFT LETTER — EDUCATIONAL, NOT FILED
A letter that reads as genuinely independent.
Start your trial