Skip to main content
Our mission

A faster way to review essays without handing judgment to AI.

Grade Harbor exists to give writing teachers a practical way to move from essay stack to reviewed feedback faster. AI prepares the draft work; teachers decide what is accurate, useful, and final.

No credit card required | You approve every grade | Plans start at $5/month

Why this exists

Writing teachers need time back and better next steps.

Essay grading asks teachers to be accurate, fast, fair, and instructional all at once. Grade Harbor focuses on the part teachers actually need help with: preparing rubric-aligned draft evidence so review can happen sooner.

Product boundary

AI can prepare feedback. It cannot know your classroom.

The system can draft a score, explain a pattern, and summarize class needs. Only a teacher understands the assignment, the student, the class context, and what feedback will actually help. Grade Harbor is built around that boundary.

Public promise

The page stays focused on what a reviewer needs to verify: teacher control, rubric-first grading, protected student data, and clear paths to support, legal, privacy, and FERPA details.

Core values

What we stand for

These principles keep the public promise practical: save teacher time, preserve teacher agency, and treat student writing as protected education data.

Draft work teachers can inspect

AI output is useful only when teachers can see how it was produced. Criterion scores, reasoning, and feedback drafts stay visible so they can be checked before approval.

Control stays with the teacher

Teachers bring the rubric, review the draft, edit the feedback, override scores, and decide when a grade is ready. The product prepares the work; the teacher owns the judgment.

Student data has one job

Student essays, scores, and skill patterns exist to support grading and instruction. Privacy, FERPA, and DPA details stay available for teachers, schools, and district reviewers.

Rubric — persuasive essay

A static rubric example with four writing criteria scored across beginning, developing, proficient, and exemplary levels.

Teacher-defined
CriterionBeginningDevelopingProficientExemplary
Thesis and argument1234
Evidence and support1234
Organization1234
Style and conventions1234
  • Beginning
  • Developing
  • Proficient
  • Exemplary

Why rubrics matter

Rubrics are more than a scoring tool.

They give teachers and students a shared map for what good writing looks like. When expectations are explicit, AI drafts can be reviewed against the same criteria a teacher would use.

Grade Harbor keeps AI-assisted scores anchored to your rubric, not a generic writing standard. That makes the draft easier to inspect, adjust, and explain before anything is final.

  • Criterion-by-criterion scoring keeps feedback actionable.
  • Consistent standards across all students in a class set.
  • AI scores against your rubric — nothing generalized.

Commitments

The principles behind the product

Human-in-the-loop, always

AI prepares a draft; the teacher decides. No grade is recorded, no feedback shared, and no action taken without explicit teacher approval. That commitment is enforced in the product — not just claimed in marketing.

Teacher agency first

Teachers set the rubric, define the criteria, override any score, and edit any feedback. The system surfaces information and recommendations — the teacher retains full authority over every instructional decision.

Student data is never sold

Student essays, scores, and skill data exist for one purpose: helping teachers improve instruction. The data is never sold, licensed, or used for advertising. Ever.

FERPA is a hard requirement

Student education records are protected by law and by design. Schools and districts can request DPA review when required, and no student data is sent to third-party services without explicit authorization.

Contact

Get in touch

Questions, partnership inquiries, or privacy/compliance requests?

Read our FERPA notice

For privacy details, see our Privacy Policy and DPA information. We do not use Student Data to train generalized AI models without explicit written authorization from the applicable school, district, or authorized educational institution.