About GradeLamp

What GradeLamp is

You paste an essay and pick what it is. GradeLamp scores each row of the rubric for that kind of essay, quotes the sentences each score is based on, and gives one concrete fix per row. College and general essays also get a plain verdict (“Not ready yet”, “Getting there”, “Ready after a polish”) before the number, because colleges don't score essays out of 30.

Right now it grades general essays, the Common App® personal essay, all 8 UC Personal Insight Questions and supplemental essays. AP® Lang, DBQ and LEQ grading opens once each passes its accuracy check against experienced AP teachers' scores. Every essay type carries a Beta tag until its accuracy has been measured and published.

Your first graded essay is free with no account, and a free account gets 3 more every 30 days. If you need more, a pass is a one-time payment that never renews (see pricing).

What it will never do

Some things are off the table at any price:

• Write, rewrite, paraphrase or “humanize” your essay. A fix is an instruction, never replacement prose.
• Detect AI writing, or promise to make anything “undetectable”.
• Predict your admissions chances, or rank you against “accepted essays”.
• Replace your teacher. A score here is a practice estimate, not an official grade.

Our academic integrity policy explains how to use AI feedback honestly, and what to tell a teacher or a college if they ask.

Why it exists

Students already ask AI chatbots and essay tools what they think of a draft. The answers are often flattering, change from one run to the next, and come with no reason you can check. Some tools rewrite the essay for you, which can get a student into real trouble. Many charge by the month and keep renewing long after the deadline has passed.

We built GradeLamp to be the opposite:

• Evidence, not vibes. Every row score points to quotes that our code has checked are really in your essay.
• Code does the arithmetic. Point caps, required elements and word limits are applied in code; the AI never adds up the score.
• The same draft gets the same result. Re-grading unchanged text returns the saved result, and small changes between drafts are shown as “about the same”, not as a gain or a loss.
• Plain privacy. Your essays are never used to train AI, there are no ads, and signed-in essays are deleted after 30 days.
• Both seasons, one account. College essays in the fall, AP essays in the spring.

How we keep it honest

Grading uses Claude by Anthropic, through Anthropic's paid API. How GradeLamp grades explains the method step by step, where it is least reliable, and how we test it.

We are measuring accuracy against experienced teachers' scores on essays written for us, and we plan to publish the results, with the numbers behind them, in January 2027. Until a result is published, the essay type keeps its Beta tag and we don't claim an accuracy figure.

Each rubric page names the qualified reviewer who checked its wording once they have signed it off, and says so plainly until then.

Who runs GradeLamp

GradeLamp is run by [Company legal name], [Registered address]. It was started by [Founder name] in [City].

You won't find testimonials, ratings or user counts here. GradeLamp is new, and we would rather show you how the grader works than tell you how popular it is.

GradeLamp is not affiliated with, or endorsed by, the College Board, The Common Application or the University of California.