Essay Grader: Feedback Before You Submit

Paste an essay between 300 and 7,000 characters and choose a level: undergraduate or graduate. An AI model reads the draft and returns two things:

• Four scores from 0 to 10 — thesis (is there a clear, arguable claim?), evidence (are claims supported?), tone (is the register academic?), and coherence (does each part follow from the last?).
• Three recommendations, each with a short title and a description of what to change.

There is no letter grade and no pass or fail. The name «essay grader» is what people search for, but what you get is structured feedback. The model does not know your instructor's rubric, the assignment prompt, or the readings, so a 7 here is not a B+ anywhere.

Example, for illustration. A master's student from Brazil pastes a 5,000-character draft arguing that cities should expand protected bike lanes, level set to graduate. A plausible result:

• Thesis 7 · Evidence 4 · Tone 6 · Coherence 5
• Support the safety claim — paragraph 2 says injuries fall after lanes are built but cites nothing; add a study or remove the claim.
• Replace absolute statements — «this always reduces traffic» overstates; hedge with «can» or «in several cities».
• Move the cost argument — it interrupts the safety section; give it its own paragraph after it.

The scores tell you where to look first; the recommendations tell you what to do there.

Writers working in a second language often have strong ideas and still lose marks in two places. Tone: English academic prose expects hedging — «suggests», «may», «appears to» — where some academic cultures prefer confident assertion, and it avoids conversational words that a direct translation can bring in. Coherence: English readers expect each paragraph to open with its point and to connect explicitly to the next, while writing traditions that build toward the point can read as wandering.

That is why tone and coherence are scored separately from evidence here. A low tone score with a high evidence score usually means the argument is fine and the phrasing needs work. Our guide to hedging in academic writing covers the most common fixes.

• It will not predict your grade or apply your course rubric.
• It will not rewrite the essay. The recommendations describe changes; you make them.
• It does not check plagiarism or tell you whether text looks AI-generated.
• It can be wrong. A model can miss a strong argument or flag something your instructor wants. Use the feedback as a second reader's opinion, not a verdict.

Getting feedback on your own draft is generally accepted in the way a writing center visit is, but some courses restrict AI tools of any kind — check your policy.

• Essay length: 300 to 7,000 characters (roughly 50 to 1,200 words). For longer papers, check one section at a time.
• Level: undergraduate or graduate.
• 10 free runs per day per IP address, no signup.

Your essay is sent to our AI provider to produce the feedback and is not stored, logged, or used for training.

If the tone score is low, run a paragraph through the sounds-translated checker to find phrasing that reads like a direct translation. If evidence is the weak axis, the paragraph structure tool helps you plan point, evidence, and analysis. You can also open the draft in the Diglot editor and work through the three recommendations there.