Features

Everything PyAutoGrader can do for your Python classroom.

Build

Assignment Creation

Create, configure, and distribute assignments from a single desktop application

🔒

Encrypted .agr Files

Assignments export as encrypted files that students cannot reverse-engineer. Solutions, expected values, and SMTP credentials are never visible to students.

📁

Multi-File Support

Target tests to specific files within an assignment. The student app auto-detects related files in the same directory when a student selects one.

📎

Bundled Data Files

Include CSV, TXT, JSON, or other data files your students' code needs. Bundled files are extracted alongside student code at grading time.

🧪

Solution Files

Upload your solution .py files for solution-comparison tests. Compare student variables, function outputs, or plot properties against your reference implementation.

📊

Partial Credit

Award proportional points when sub-checks partially pass. A test with 10 points and 5 sub-checks awards 2 points per passing sub-check.

💬

Custom Feedback

Configure pass/fail messages and hints for each test. Students see targeted guidance on failure, helping them learn from mistakes without seeing the answer.

Test

60+ Built-In Test Types

From basic variable checks to cyclomatic complexity analysis - test what matters

Variables & Data

10 types

variable_valueCheck a variable equals an expected value with optional float tolerance

variable_typeVerify variable type (int, float, str, list, dict, tuple, etc.)

list_equalsCompare lists with optional order checking and float tolerance

array_equalsNumPy array comparison with element-wise tolerance

array_sizeCheck array dimensions (min, max, or exact size)

array_values_in_rangeVerify all array values fall within numeric bounds

dict_equalsDictionary comparison with float tolerance

compare_solutionCompare student variables against your solution file

check_relationshipVerify mathematical relationships between variables (e.g., y > 2*x)

dataframe_columnsCheck a pandas DataFrame has expected columns

Functions

6 types

function_existsVerify a function is defined in student code

function_calledCheck a function was called during execution

function_not_calledVerify a function was NOT called (detect unauthorized helpers)

test_function_solutionTest student function against your solution with specific inputs

function_signatureCheck function parameters, parameter count, and defaults

return_typeVerify a function returns the expected type

Classes

3 types

class_existsVerify a class is defined

class_has_methodCheck a class has a specific method

class_has_attributeCheck a class has a specific attribute

Control Flow

5 types

for_loop_usedDetect presence of for loops

while_loop_usedDetect presence of while loops

if_statement_usedDetect presence of if statements

loop_iterationsVerify exact loop iteration count via a tracked variable

exception_raisedCheck that code raises a specific exception type

Code Patterns

5 types

code_containsSearch for a phrase in source code (case-sensitive option)

code_not_containsVerify a phrase is absent from source code

operator_usedDetect specific operators (+=, **, //, etc.)

stdout_containsVerify printed output contains expected text

import_usedVerify a specific module was imported

Matplotlib Plots

13 types

plot_createdVerify a matplotlib plot was generated

plot_propertiesCheck title, xlabel, ylabel, legend, and grid at once

plot_has_xlabel / plot_has_ylabel / plot_has_titleIndividual label and title checks

plot_data_lengthVerify number of data points per line (min, max, or exact)

plot_line_styleCheck line style by index (solid, dashed, dotted, dashdot)

plot_has_line_styleCheck if any line uses a given style

plot_line_widthCheck line width with tolerance

plot_marker_sizeCheck marker size with tolerance

compare_plot_solutionCompare against your solution plot (color, style, width, markers)

check_multiple_linesVerify a plot has at least N lines

check_exact_linesVerify a plot has exactly N lines

check_function_any_lineCheck if any line matches a mathematical function

Code Quality

15 types

naming_conventionEnforce snake_case, camelCase, PascalCase, or custom regex

no_single_letter_varsBlock single-character names (with exceptions like i, j, k)

line_lengthMaximum characters per line with optional tolerance

code_spacingWhitespace around operators, after commas/colons, no trailing spaces

has_docstringsRequire docstrings on all functions, public-only, or specific functions

has_module_docstringRequire a file-level docstring

min_comment_densityMinimum percentage of commented lines

max_function_lengthMaximum lines per function (optionally including docstrings)

max_nesting_depthMaximum nesting level for control structures

cyclomatic_complexityMcCabe complexity threshold per function

no_wildcard_importsBlock from x import *

imports_at_topRequire all imports at the top of the file

no_global_variablesDisallow global-scope variable assignments (constants exempted)

Performance

4 types

performanceFunction must complete within a time limit

performance_scriptFull script must complete within a time limit

performance_compare_functionCompare function speed to your solution (averaged over multiple runs)

performance_compare_scriptCompare full script speed to your solution

File Naming

2 types

file_nameCheck submission filename (exact match, prefix, suffix, contains)

file_naming_conventionEnforce filename conventions (snake_case, kebab-case, etc.)

Collect

Collect Results Your Way

Five channels, any combination. Use what fits your workflow.

📄 PDF Export

Students export a formatted PDF with scores, per-test results, and their code. They upload it to your LMS like any other assignment.

Can be required by instructor

📬 Submission Server

Full results and student code sent to your self-hosted server. Dashboard with analytics, plagiarism detection, code search, and Canvas export.

Can be required by instructor

📧 Email

Results emailed to you automatically on each submission. Includes student info, scores, per-test results, and the .py file as an attachment.

Silent notification - cannot be required

📊 Google Sheets

Scores posted to your Google Sheet via an Apps Script webhook. Each submission adds a row with student info, scores, and per-test pass/fail.

Silent notification - cannot be required

🔗 Webhook

POST a JSON payload to any URL on each submission. Build custom integrations with your existing infrastructure.

Silent notification - cannot be required

All channels operate independently. If one fails (no internet), submissions are queued locally and retried automatically.

Grade

Batch Grading

Grade a folder of student submissions and export a CSV gradebook

📥

1. Download

Collect .py files from your LMS

⚡

2. Grade

Load the folder in the instructor app and run batch grading

📋

3. Export

Get a CSV with per-test scores, totals, and percentages

Student names are extracted automatically from file comments (# Name: John Smith), filename patterns (Smith_John_hw1.py), or the filename itself. Export as standard CSV or Canvas-compatible format for direct LMS import.

🛡️

Import Safety Scanning

Before executing each file, the grader uses Python AST analysis to scan for dangerous imports across 15 blocked modules - including subprocess, socket, shutil, and more. If one is detected, a modal lets you allow or deny execution before proceeding.

Identity

Roster Management

Optional student identity validation with three roster sources

📋

Bundled CSV

Embed a roster directly in the .agr file. Best for small classes and offline-only environments. Students select their name from a validated list.

☁️

Google Sheets

Fetch the roster live from a published Google Sheet. Late adds take effect immediately - no .agr redistribution needed.

🖥️

Server API

Fetch from your submission server's REST endpoint. Manage rosters through the dashboard with CSV upload, manual add, and duplicate detection.

Combine a remote source with a bundled CSV for offline fallback. Three name modes: freeform entry (default), strict username validation, or flexible username-or-name.

Manage

Self-Hosted Submission Server

A full-featured Flask server you run on your own infrastructure

📊

Analytics Dashboard

Score distributions, submission trends, assignment difficulty analysis, student progress tracking, and at-risk student identification. Filter by course, section, or semester.

🔍

Plagiarism Detection

Pairwise code similarity analysis with four metrics: code similarity, comment similarity, combined score, and comment-to-code ratio. Side-by-side diff view with syntax highlighting.

🔎

Code Search

Search across all submitted code for specific patterns or keywords. Detect AI-generated code patterns, verify required constructs, or investigate academic integrity concerns.

📅

Assignment Deadlines

Set per-assignment cutoff dates with per-student overrides for accommodations. Late submissions flagged with badges. Exports include an option to exclude late work.

📤

Canvas Export

Export grades as Canvas-compatible CSV for direct LMS import. One row per student, one column per assignment. Respects view mode and filter settings.

⚙️

Zero-Config Setup

SQLite database requires no setup. Download the executable, set a password, and run. Optional PostgreSQL for high-concurrency deployments. HTTPS via Caddy reverse proxy.

Learn more about the server

Trust

FERPA-Friendly by Design

Your data stays on your infrastructure. Always.

🖥️

Local Grading

Student code executes on the student's own machine. No cloud servers process student work.

🔒

Encrypted Distribution

Solutions, expected values, and SMTP credentials are encrypted inside .agr files. Students cannot extract them.

🏛️

Self-Hosted Infrastructure

The submission server runs on your machines. Student data never passes through PyAutoGrader's infrastructure.

📝

No Third-Party Agreements

No data processing agreements, no vendor contracts, no per-student fees. You control the entire pipeline.

Missing a test type?

We actively add new test types based on instructor feedback.

Request a Test Type Download PyAutoGrader