Automatic Grading Platform

Evaluation
That Teaches.

Built for educators and institutions

Evaluating code is usually slow and doesn't scale. With CodeSpell, you can automate your entire grading workflow — tests, quality, and code style — and get results in under 30 seconds.

binary_search.pyutils.py
1# returns index of target, or -1
2
3def binary_search(arr, target):
4 left, right = 0, len(arr) - 1
5 while left <= right:
mid = left + right // 2
7 if arr[mid] == target:
8 return mid
9 elif arr[mid] < target:
10 left = mid + 1
11 else:
12 right = mid - 1
13 return -1
PythonLn 6, Col 281 error · 2 warnings
Overall Score
92%Passed
Unit Tests
8 passed0 failed
Correctness95%
Code Quality88%
Complexity72%

By the numbers

Built to save time
at every scale.

80%
Grading time savedvs. hours of manual review per assignment
01
5
Languages supportedPython, Java, JavaScript, C, C++ · React & Node.js soon
02
15 min
Average setup timefrom sign-up to your first graded submission
03
1500+
Submissions gradedand counting, across all institutions in early access
04

Solution

Automatic Code Evaluation
That Actually Works

Binary Search — CS 101

Python · Due in 3 days

Draft

Rubric

Correctness40%
Code Quality30%
Complexity30%
AI generating assignment content…

01

Create

Build your assignment from scratch or let AI generate content from your course materials. Set the rubric, define the criteria, and publish to your students — all in one place.

Outcome: READY IN MINUTES
Py

binary_search.py

Python · 2.4 KB

JS

sorting_utils.js

JavaScript · 1.8 KB

C++

linked_list.cpp

C++ · 3.1 KB

JavaC++JSGo

02

Submit

Students upload code directly or write it in the integrated IDE — no distractions, no additional setup. Works with any language, any format, any complexity.

Outcome: INSTANT ANALYSIS
Py

binary_search.py

Python · 2.4 KB

Final ScorePassed
92%
Code Quality
88%
Complexity
72%
Security
93%
Graded in 11s8/8 tests passed

03

Grade

We run your tests, apply your rubrics, and evaluate the code's performance, quality, and security — all at once. Every submission gets detailed, consistent results in under 30 seconds.

Outcome: DETAILED RESULTS

Platform

The complete grading
workflow, in one place.

01Assignment creation

Build your assignment. Share instantly.

Write your own assignment content or let AI generate it from your course materials. Attach a rubric, set the deadline, and publish directly to your students — no external tools needed.

Title

Binary Search Implementation

Instructions

AI generated

Implement a binary_search(arr, target) function that returns the index of the target in a sorted list, or -1 if not found.

Your solution must run in O(log n) time and handle empty arrays and duplicate values.

Rubric

100 pts total
Correctness40%

All test cases pass, edge cases handled

Code Quality30%

Readable, well-named, no dead code

Complexity30%

Optimal time/space complexity

Language

PythonJavaC++

Due date

Apr 25, 2026

in 6 days

Assigned to

CS 101

32 students

2 attachments
02Inline feedback

Issues visible the moment they run.

Every problem is ranked by severity, linked to the exact line, and explained in plain language — no manual code review needed.

main.pyutils.py
1# returns index of target, or -1 if not found
2
3def binary_search(arr, target):
4 left, right = 0, len(arr) - 1
5 while left <= right:
mid = left + right // 2
7 if arr[mid] == target:
8 return mid
9 elif arr[mid] < target:
10 left = mid + 1
11 else:
12 right = mid - 1
13 return -1
PythonLn 6, Col 281 error · 2 warnings
Top Issues3 found
highOperator precedence error

mid = left + right // 2 should be (left + right) // 2

main.py:4
mediumHigh cyclomatic complexity

Function binary_search has complexity score of 5

main.py
lowMissing type annotations

Add type hints to parameters and return value

main.py:1
03Automated grading

Your rubric. Applied to every submission.

Define your criteria once. Unit tests, complexity, security, and style are all weighted and scored consistently — down to the decimal.

89%overall score
A−
24 passed3 failed
Unit Tests
100%35% wt.
Code Quality
88%25% wt.
Complexity
72%25% wt.
Security
93%15% wt.
04Class analytics

Know who's struggling before the deadline.

Submission trends, pass rates, and activity across every student in one view. Act early, not after grades are due.

24

Submissions

87%

Pass rate

91.2

Avg. score

Activity

Last 10 weeks

Less
More

Recent

Binary Search

Python · 2h ago

A−

Linked List

Java · 1d ago

B+

Stack Impl.

C++ · 3d ago

F

Merge Sort

JavaScript · 5d ago

A

How it works

We automate
grading the way
code should work.

01

Create your assignment

Write your own content or let AI generate it from your course materials. Define the problem, attach a rubric, and publish directly to your students — all from within CodeSpell.

02

Applies your rubric

We grade according to YOUR criteria with exact weighting. AI applies your rubric — not a generic standard — and generates specific, line-level feedback for each criterion the student missed.

03

Runs your test suite

We execute your custom tests in isolated containers against every submission — consistent, safe, and reproducible regardless of language or complexity.

04

Checks algorithm efficiency

We analyze time complexity, memory usage, code structure, and adherence to best practices for the specific assignment type.

05

Detects plagiarismComing soon

We compare submissions across your current class and prior semesters, flagging structural similarities and shared code patterns.

06

Integrate your workflow

Using our REST API, you can create content, access detailed evaluation results, and integrate CodeSpell's grading into your existing tools and workflow.

07

Syncs to your LMSComing soon

Grades and feedback will sync directly to Canvas, Moodle, or Blackboard the moment evaluation completes. Students see results immediately.

Use Cases

Built for every
kind of evaluation.

01University CS Departments

Grade 300 submissions in the time it takes to grade 50.

Weekly assignments across multiple sections mean hundreds of submissions per deadline. CodeSpell runs your rubric against every one — consistently, instantly, without TAs burning out on repetition.

PythonJavaC++LMS sync · soon

80%

reduction in grading time

02Secondary Schools

Grade more often. Teach better. No CS degree required.

Secondary school teachers shouldn't need to be programming experts to give students meaningful feedback on code. CodeSpell handles the evaluation — so teachers can assign more exercises, spot struggles earlier, and focus on what they actually know: teaching.

Beginner-friendlyPythonJavaScriptAny level

2×

more frequent evaluations

03Coding Bootcamps

Students get feedback before the next lecture starts.

Bootcamp students submit daily. Waiting 48 hours for feedback breaks their momentum. CodeSpell returns detailed results in under 30 seconds — so instructors coach instead of correct.

JavaScriptPythonJavaInstant results

< 30s

average feedback turnaround

04Corporate L&D

Assess real coding skills — not just multiple choice.

Onboarding and upskilling programs need objective evidence of technical ability. CodeSpell evaluates working code against your exact criteria, so hiring and promotion decisions are grounded in data.

Custom rubricsMulti-languageReporting

100%

objective, rubric-based scoring

FAQ

Common questions

Start grading smarter
today.

Free during our early access pilot. No credit card, ever.

Questions? Contact us →