Your application, read against thousands of real admitted essays.
ChatGPT was trained on the open internet — it has never seen a real admissions file. We trained on thousands of real admitted, waitlisted, and rejected packages, each with the actual outcome attached. Upload yours, see exactly what's quietly costing you, and re-run after every revision until it's ready.
After that: $0.99 a read, or the whole season for $199.
Three places you could get your package read. Only one has the files.
- Trained on the open internet
- Has never seen a real admitted Yale package
- Same generic answer for every applicant
- “Be authentic. Show passion.”
- Thousands of real packages — multiple firms, growing every cycle
- Every file has the real admit / waitlist / reject outcome attached
- Compares your specific profile to specific real outcomes
- First read free — no incentive to keep you around
- A few hundred packages across an entire career
- Paid more the longer you keep booking sessions
- Won't tell you a reach is honestly out of reach
- Sample size of one career — not thousands of cycles
One upload. A long, honest read. Then iterate until it's ready.
We can't change who you are — and wouldn't. We change how you organize and tell what you already have, so the readers at your top schools actually see it. Most applicants are losing on storytelling, not credentials.
Common App essay, supplements, activities list, transcript notes. PDF, Word, or paste-in — the AI reads everything in one pass.
A calibrated admission probability, school-specific prompt fit, strengths and weaknesses named specifically, and line-level rewrites. Roughly 4,000–6,000 words about your package, not applicants in general.
Revise and run it again. Each new read compares against your last draft — what landed, what still needs work, exactly what to attack next.
One read is a second opinion. The loop is where applications actually get better. Every run remembers your previous drafts and tells you which suggestions you acted on, which still need work, and exactly what to attack next — across every essay and every school on your list.
Same student, same credentials — a sharper story, told the way the readers at that school actually reward.
Illustrative. Real movement depends on your package.
- An admission probability calibrated to that school's actual acceptance rate
- School-specific prompt-fit scoring against the current cycle's essay questions
- Strengths and weaknesses, named specifically, not in clichés
- Essay-by-essay scoring with line-level rewrites and a draft rewrite in your voice
- Activities list reordering — which to feature, which to cut, where to add evidence
- Concrete things to ADD if true of you (jobs, awards, family context, supplements you skipped)
- An iterate loop: edit, re-run, compare — until the report stops finding leverage
- Empty “you have such a unique voice!” affirmation
- A consultant trying to upsell you a yearlong package
- Generic, school-agnostic advice you could find on Reddit
- Anything that sounds like it came from a brochure
Yes — it will write your essays if you ask. Don't let it.
On any run you can ask for a full drafted rewrite — your story, your details, your voice. It's genuinely good. That's exactly why you shouldn't submit it: selective colleges now run essays through AI-detection software, and a pasted-in AI essay is the fastest way to get an otherwise-strong application flagged.
Use the draft the way strong writers use an editor: to break writer's block, to see the shape of a sharper essay, to steal the structure — never the sentences. Then write it yourself. The read tells you what to fix; the words have to be yours.
- Break out of writer's block with a draft in your voice
- See what a tighter structure looks like for YOUR story
- Compare your draft against what actually got people in
- Rewrite it yourself, then re-run and watch the score move
- Paste an AI-drafted sentence into your Common App
- Submit anything you couldn't have written yourself
Five real reads. Yours will be this specific.
Not a grade and a pep talk — a calibrated probability, your essays scored line by line, and an exact list of what to fix tonight. Five applicants, five top schools, the same brutal honesty. Each runs ~4,000 words.
Five real-style reads · different applicants, different schools · the same brutal honesty. Tap any card.
The whole point of the read is this list — specific, do-it-tonight moves, ranked by how much they actually change the outcome. Not “be more authentic.” This:
Re-frame the activities and supplement around the translator — not a separate pre-law plan.
Trade brochure language for one seminar, one professor, one argument you'd push back on.
Delete the 'taught me about identity' line and the brochure conclusion. Five minutes.
Below: the full read this came from — probability, scorecard, essay-by-essay line edits, and drafted rewrites you can start from.
Strong Common App, weaker supplements.
Real proof, but two entries tell one story.
Two competing stories, not one.
Top decile of Yale's pool.
'Why Yale' reads like the brochure.
You read like a humanities-strong applicant in the top 12% of what Yale will see this year. Your Common App essay does one rare thing well: it trusts the reader. But your supplements undo that work — the 'Why Yale' supplement names exactly one specific thing (Directed Studies) and then drifts into Yale's admissions brochure language. That's a tell, and the reader will catch it.
Here is what a committee actually sees: a writer who can clearly write, attached to a package that hasn't decided what it's about. The single biggest lever is not another activity or a higher score — it's collapsing your two narratives into one. Do that and this moves from 'strong but forgettable' to 'the translator,' which is a person they remember in committee.
The honest ceiling: even executed perfectly, Yale is a reach at a 3.7% base rate. A clean, single-story package puts you at the top of the reach band — not into target. We'd rather tell you that now than sell you a year of polish that can't change the math.
You're closest to two admitted Yale archetypes: the 'quiet humanist' (admit, 2023) and the 'arts + intellectual' cellist who named a real disagreement with a Yale professor in her supplement (admit, 2024). The cellist's essay-to-supplement coherence is exactly what you're missing. The closest reject in our files had your strengths and your problem: a beautiful main essay bolted onto supplements that could have been submitted to any school.
- Voice in the Common AppRestrained, sentence-level writer. Rare in the pile, and it can't be taught in a year.
- Activities have proof, not adjectivesNumbers and dates exist where they should; you show rather than claim.
- One genuine intellectual interestTranslation work named with real depth — this is your throughline if you choose it.
- 'Why Yale' reads as brochure-fluent'Vibrant intellectual community' is what readers filter out, not in.
- Activities #2 and #6 tell the same storyPromote #9 instead — the only non-academic risk in the file.
- The supplement undermines the essayIt says out loud what the main essay wisely refused to say.
Your package is trying to tell two stories at once, and they're fighting each other for the reader's attention.
A translator (the essay) and a future lawyer (activities, supplements). The lawyer story is one Yale reads 800 times a cycle.
Build everything around the translator — the rarer, stronger story. Re-frame the activities as evidence of it, not a separate career plan.
The opening image (a misspelling in your grandmother's 1962 letter) is the kind of detail Yale readers underline. You let the reader do the inferring — most applicants don't — and you never explain the metaphor, which is exactly right.
Paragraph 3 lapses into 'this taught me about identity,' and the conclusion reaches for a neat bow the rest of the essay was too smart to need. You stop trusting the reader right when it counts.
“…and that's when I realized this taught me about identity.”
Issue: You name the lesson the image already delivered. It deflates it.
Fix: Cut the sentence. End the paragraph on the letter itself.
“I will carry this lesson with me to college and beyond.”
Issue: A brochure sentence in an essay that earned a better ending.
Fix: Return to a concrete object — the letter, refolded — and stop.
“The letter is dated August 4, 1962. The word 'recieved' is misspelled in my grandmother's otherwise impeccable hand, and for forty years no one corrected it.”
Now: “I will carry this lesson with me to college and beyond.”
“I refolded the letter along its forty-year creases and put it back. I still haven't corrected the spelling. Some things you keep exactly as they were handed to you.”
The 'Why Yale' supplement is the weakest piece and the easiest win. Ask and we'll draft a full rewrite in your voice to start from — then name one seminar and one professor whose argument you'd push back on, and run it back through. We'll tell you if it landed.
Re-frame the activities and supplement as evidence of the translator, not a separate pre-law plan.
Replace brochure language with one seminar, one professor, one argument you'd push back on.
Delete the 'taught me about identity' line and the brochure conclusion.
This is run 1. Make the edits, re-upload, and run it again — each time we tell you whether your probability moved, which fixes landed, and where you can still do better. It rarely jumps on one pass. It climbs over a season of revisions:
- Single translator narrative now carries the package.
- 'Why Yale' names a real professor and argument.
- Activity #9 still buried at the bottom.
- Second supplement still hedges its one risk.
Three single reads barely covers one school's first draft. Unlimited is the version built for the loop: re-read after every revision, run your package against every college, customized per school, from now through decision day. One price for the whole season — and if you ever reach the cap as a real applicant, we just add more turns.
Illustrative numbers — real movement depends on your package, and we never inflate the probability.
Use it for every college. Each read is tailored to that school's actual current prompts.
We've verified the supplemental essay questions each top school asks and what each is known to publicly value. Your Yale read scores against Yale's actual prompts. Your Stanford read scores against Stanford's. We tell you when a generic answer is missing a real anchor for that specific school.
- · Harvard's five required short answers, individually
- · Yale's 125-word Why answers and 400-word essay
- · MIT's pleasure-prompt and collaboration prompt
- · Stanford's roommate letter and contribution essay
- · Penn's school-of-choice fit answers
- · UChicago's yearly extended essay
The archive is the product.
The admissions industry runs on a single trick: convince a 17-year-old and their parents that the process is so opaque only a $5,000–$50,000 consultant can decode it. Consultants read your package, call it "promising," and propose another year of sessions.
Over years of partnerships with multiple consulting firms — and thousands of applicants who shared their packages directly — we collected the complete files of thousands of admitted, waitlisted, and rejected applicants, every one with the real outcome attached. Every file anonymized, with explicit consent, from the past three cycles only. This is the dataset no general AI has, and the only way to read your package against the real ones that won and lost.
A $499 service. Now $199 for the whole season.
Reading your package against thousands of real admitted, waitlisted, and rejected files — an archive built over years of partnerships with admissions consulting firms — is genuinely a $499 service; a single consultant session costs more. While we grow, the whole season is $199. And the first read costs nothing at all.
Code FIRSTREAD — first 300 applicants. Put your real package in, read the full report, then decide.
- First read free with code FIRSTREAD (first 300 applicants)
- Calibrated admission probability against base rate
- School-specific prompt fit against current prompts
- Strengths, weaknesses, line-level essay rewrites
- Things to add if true of you (jobs, awards, context)
- Up to 3 reads per email — then switch to Unlimited
- Unlimited reads from now through decisions
- Re-read after every revision, as many times as it takes
- Run against every college on your list — customized per school
- Everything in the single read, every time
- Hit the cap as a real applicant? Email us, we add more turns
On Unlimited: you're buying the whole admissions season — from now through decisions. It's metered at 499 reads, far more than any student needs, purely to stop professional resellers. We will never cut a real applicant off mid-revision. Admit Files (full anonymized admitted packages) is paused while we focus on the read — it'll be back.
Questions, answered honestly.
How is this not just ChatGPT with a wrapper?+
Because ChatGPT, Claude, and Gemini have never seen a real admitted package. They were trained on the open internet, and the open internet does not contain last year's admitted Yale essays. We have those — thousands of them, with the real admit / waitlist / reject outcome attached to every single one — built up through partnerships with multiple admissions consulting firms and thousands of applicants who've shared their packages with us directly. The archive grows every cycle. The AI model is the engine. The archive is the fuel.
What does the iterate loop actually do?+
You upload your package, get the long structured report, then revise an essay (or ask the AI to draft a rewrite in your voice). Re-run the read and it tells you what landed compared to your previous draft, where you've improved, and where there's still leverage. We keep doing this until the report stops finding meaningful things to fix. Each iteration is a fresh evaluation; the AI remembers your previous drafts.
Will you write my essay for me?+
It will, if you ask — full drafts and rewrites, built from your own story and details. And we're telling you plainly: don't submit them. Selective colleges run essays through AI-detection software, and a pasted-in AI essay is the fastest way to get your application flagged. Use a draft to break writer's block or to see the shape of a stronger essay, then write it yourself, in your own words. Every rewrite we produce carries that warning, every time.
Can I use it for multiple colleges?+
Yes — and you should. Each read is customized per school. Your Yale read scores against Yale's actual current prompts; your Stanford read scores against Stanford's. We've verified the current essay questions each top school asks and what each is publicly known to value. Run your same uploads against every college on your list.
What's the FIRSTREAD promo code?+
Your first read is free with code FIRSTREAD — capped at the first 300 applicants to use it. After that, $0.99 per read, up to three reads per email. Iterating past that? Switch to Unlimited.
What exactly am I buying with Unlimited?+
One full admissions season of reads for $199 — from now through decision day. Run your package as many times as it takes, across every college on your list, customized per school. It's metered at 499 reads only to keep professional resellers out; that's far more than any real student needs to iterate freely. If you're a genuine applicant who somehow reaches the cap mid-season, email us and we'll add more turns. We will never cut you off in the middle of revising.
Is the probability a real number, or just made up?+
It's calibrated against the school's actual acceptance rate and the corpus archetypes. A typical strong applicant at Harvard won't read 40% — it'll read 5–10%. We will not inflate numbers; the whole product is worthless if we do.
What if I have nothing yet — no essays, just an idea?+
Then this isn't for you yet. Come back when you have a Common App draft and an activities list. We grade what's on the page.
What about the Admit Files product?+
Paused for now. We're focused on the iterate-and-improve service first. The archive that powers the reads is the same archive that would power the Admit Files; both products will return.
What happens to my essays after I upload them?+
Stored encrypted in our Postgres database, used to generate your report, and never sent to any model provider for training. You can delete the evaluation any time. Iteration drafts (revised essays + AI-generated rewrites) are kept alongside the evaluation row for the duration of your iterate loop.
Where did the archive come from? Is this even legal?+
We've spent years building it. Multiple admissions consulting firms partner with us, and thousands of applicants have shared their packages with us directly. Every applicant in the archive consented to anonymized use, every file is stripped of personally identifying information, and we only keep cycles from the past three years. The archive grows with every application season.
Upload your package. The first read is on us.
Trained on thousands of real admitted, waitlisted, and rejected packages — files no other AI has. Customized per college. Free with code FIRSTREAD, then iterate all season for $199.