How it works For students The engine Pricing Security
Sign in Start free
New Google Classroom grade posting available; SIS connections are partner-gated

Grade faster. Keep the final say.

Give teachers their evenings and weekends back by taking over the grading load: reading student work, scoring it against the rubric, and drafting feedback, while the teacher stays the final authority on every grade.

Start free, no credit card See how it works
100 free grades/mo You're always the final authority FERPA-aware by design
✦ Period 3 · Calc · Quiz 7
Grading 28 submissions
Student work
28 scans
Photos · PDFs · Classroom
Confidence routing
✓
Auto-post
≥ 95% confident
0
≈
Bulk-approve
80-95%
0
!
You review
< 80%
0
Gradebook
Teacher-approved results
A. Nguyen19/20
M. Patel18/20
J. Rivera20/20
S. Cohen17/20
How it works

You're the referee, not the data-entry clerk.

Skyline grades everything, then routes each result by how sure it is. The more confident it is, the less it asks of you.

Auto-post

When Skyline is highly confident and you explicitly enable auto-post, it can finalize eligible grades automatically. Gradebook posting still requires a supported, connected integration.

  • Conservative by default. Auto-post stays off until you enable it in your teacher settings.
  • You're notified after the fact, with a full 48-hour window to override any grade.
  • Every auto-posted grade is logged with its confidence and reasoning.

You stay in control. Auto-post never runs without your say-so, and you can claw any grade back inside 48 hours. This is automation that answers to you, never the other way around.

Bulk-approve

In the 80-95% range, Skyline has a strong answer but wants a second set of eyes. It batches these so you can scan and approve a whole stack at once.

  • Grouped by assignment and by the rubric criterion in question.
  • Skim the proposed grades and approve the whole batch in one click, or open any single item you want a closer look at.
  • Adjust one score and Skyline learns your calibration for next time.

Minutes, not hours. A class set that used to take an afternoon becomes a quick read-through. You're confirming the call, not redoing the work.

Individual review

Below 80%, maybe messy handwriting or a genuinely creative approach, Skyline won't guess. It lays out everything you need to make the call in seconds.

  • The original work, the proposed score, and the AI's reasoning, side by side.
  • Per-criterion breakdown so you see exactly where the uncertainty sits.
  • Accept, adjust, or rewrite. Your call posts to the gradebook with one click.

This is where you matter most. Skyline clears the routine so your judgment goes to the handful of answers that actually need a teacher.

For students too

Students can turn in work without another account.

Share one class code or link. Students choose an open assignment, upload a photo or PDF, and can confirm that their work reached the grading workflow.

Join with a class code

No account, no app. Students open your class link or type the code, pick their name, and upload a photo or PDF.

Clear submission status

Students can confirm whether work is grading, with the teacher, returned, or needs another upload. Unreleased scores stay private.

You stay in control

Students see status only. You review and release grades and feedback through the teacher workflow.

Student? Turn in your work
Under the hood

Most tools make one LLM call. Skyline runs a grading engine.

Three layers work together so a grade is never a guess: verified where it can be, reasoned where it can't, and honest about how sure it is.

Math Physics Chemistry Biology English essays History DBQs

Subject-specific evaluators, because a math grader should work nothing like an essay grader. The engine is hard to copy precisely because it isn't a prompt.

Symbolic verification SymPy

For math and quantitative science, answers are checked symbolically, so 2x·sin(x)+C and sin(x)·2x+C are the same answer. No model hallucination, just math.

01

LLM reasoning over your rubric

For method, justification, and prose, an LLM evaluates the work against your structured rubric, one criterion at a time, and writes feedback that points back to the rubric instead of vague platitudes.

02

Confidence estimator

A separate layer scores how sure the engine is, and that number decides the routing tier. Uncertainty is surfaced, never buried under a confident-sounding grade.

03
Closing the loop

Approved grades can move into the tools your school uses.

Google Classroom grade posting is available after teacher review. Aeries, PowerSchool, Canvas, and other district connections remain pilot, partner, or roadmap work until verified for a school.

LMS

Where work comes in

Supported
connections
Current and planned

Gradebook connections

ai_proposal: 19/20 confidence: 0.97 teacher_action: approved rubric_version: v3 destination: classroom ts: 2026-05-29T16:04Z

Connection availability depends on provider approval, district setup, assignment mapping, and successful live verification.

Trust & audit

Every grade is accountable. You are always the final word.

Grading touches student records, so Skyline is built to be inspectable, privacy-first, and conservative about what leaves your classroom.

A complete audit log

Every grade records the AI proposal, confidence, reasoning, your action, the timestamp, and the rubric version. Reconstruct any decision, any time.

Teacher is final authority

Skyline proposes; you decide. Override windows, individual review, and per-assignment auto-post controls all keep the human in charge.

No student PII to LLMs

Names and identifying details are stripped before any model sees the work. FERPA- and COPPA-aware from day one.

No training on student work

Student submissions are never used to train models without explicit consent. SOC 2 is on our roadmap.

audit-trail · submission #4471
16:04:02.118ai_proposal 19/20 · per-criterion [5,5,4,5] · confidence 0.97
16:04:02.119route → auto-post (≥ 0.95)
16:04:03.402teacher_approved final score saved · rubric v3
09:14:22.001teacher_review opened · A. Chen
09:14:48.553teacher_action confirmed · no change
Pricing

Start free. Add your department when you're ready.

Individual teachers can start free. Paid and institutional plans expand usage, controls, and supported integration onboarding.

A Standard grade uses 1 credit; a Thorough grade, a deeper pass for essays and tricky work, uses 4. Credits refresh monthly. Paid plans are in early access while billing comes online.

Free
For the individual teacher trying it on their own class.
$0
No credit card, ever.
Start free
  • 100 Standard grades / month
  • Core grading + drafted feedback
  • Google Classroom only
  • Confidence routing & audit log
Teacher
For the teacher who grades across every section, all term.
$12/ month
Early access · billed monthly at launch.
Join early access
  • 600 grading credits every month
  • Thorough grading for essays & edge cases
  • Full rubric library
  • Supported LMS workflows
  • Advanced feedback controls
  • No SIS write-back
Teacher Pro
For heavy graders: several sections, long essays, all term.
$29/ month
Early access.
Join early access
  • 1,500 grading credits every month
  • Everything in Teacher
  • Thorough grading at scale
School & District
For teams standardizing grading with shared rubrics and calibration.
From $6/ teacher / mo
Pooled credits · 5+ teachers · billed annually.
Contact sales
  • Pooled grading credits across the team
  • Everything in Teacher Pro
  • Shared rubric library & calibration analytics
  • Admin controls & district audit logs
  • Integration onboarding (availability verified per provider)
District integration support starts here

Get your evenings back this week.

Start grading your next set free. No credit card, no setup call. You stay in control of every grade, and Skyline just clears the routine.

Start free, no credit card See how it works