Project Info
This project did not submit a demo video on Devpost.
Inspiration
As you may or may know as students or professionals in STEM, the jump in math education to primarily problem solving based learning to proof writing is a difficult switch for many students and professors alike. Based on our experiences taking and grading proof-focused courses at our universities and the lack of targeted practice available for learning this essential mathematical skill, we created Proof Buster. Description Proof Buster is an educational tool which provides students with an assisted gamified approach to proof writing inspired by similar platforms for programming and language learning. Through Proof Buster, students select a conjecture to prove and learn by doing. Unlike traditional tutors and teaching assistants who must carefully read through long paragraphs to identify errors, our AI tutoring agent which leverages Gemini’s speed at identifying common errors to provide accurate and immediate feedback to the user. Through Gemini’s ability to add additional model tuning and directly call functions within our project, we are able to leverage AI’s data processing strength to both save hours of labor for instructional teams and provide a much more intuitive and engaging learning experience for students. Development Proof Buster uses a standard frontend and a node.js based server for its backend. The backend utilizes Google's Gemini API to train the model on conjectures and ready-made proofs to fine tune it to our project needs and avoid errors with its assessment of students' proofs as well as Gemini's function calling to get non-String values for use throughout our project. Despite some unfamiliarity with JavaScript and a team size of just 2, we were able to create a fully functional website that accomplished all the basic features which we sought out to perform. Looking Forward Proof Buster is not just a throwaway project; it was specifically designed to help us with proof-based math courses and their education. With small tweaks and additions such as a web-based server, user login, conjecture submission, and regularly added conjectures, we believe Proof Buster will be instrumental in helping teach students in classes we grade or take ourselves.
Analysis
View
Metric
- 3
Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.
Technology
- CSSIn code
- ExpressIn code
- Google GeminiIn code
- HTMLIn code
- JavaScriptIn code
- Node.jsClaimed
5 of 6 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.
AI coding agents
No AI coding agent signals were found in this repository.
Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.
Codebase size
Source size
19 KB
Source files
6
Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.
Repository
eliaswuberkeley/proofbuster.github.io
13 files · 316 KB · @ cfa0000
Structure
Application logic
2 files · 15%Domain rules, services and shared utilities.
Supporting
Layers are inferred from where files sit in the tree, not from reading the code. A project that names its directories unconventionally will read oddly here — open the file browser to check anything the diagram implies.
Languages
- JavaScript49%
- HTML29%
- CSS22%
- Markdown0%
Share of indexed source by file size. Binary and vendored files are excluded.
Dependencies
package.json
npm · 5- @google/generative-ai
- body-parser
- dotenv
- express
- showdown
Declared in the repository’s manifests at the indexed commit. A declared package is not proof it is used, and runtime dependencies are listed first.
Feature verification
AI tutoring agent powered by GeminiVerified
AI tutoring agent which leverages Gemini's speed at identifying common errors
Claimed on Devposthigh confidenceserver.js:17— GoogleGenerativeAI client initialized with gemini-1.5-flash model and used across all three endpoints
Conjecture selection to proveVerified
Students select a conjecture to prove and learn by doing
Claimed on Devposthigh confidencepublic/script.js:6— Loads problems.json and populates a dropdown of conjectures for the user to selectpublic/problems.json:1— List of predefined conjectures with difficulty and tags
Custom conjecture submission and validationVerified
Small tweaks such as conjecture submission (mentioned as a current capability via the Custom UI, distinct from the future-work list)
Claimed on Devpostmedium confidenceserver.js:53— /check-custom endpoint asks Gemini whether a user-submitted conjecture is a true/valid conjecturepublic/script.js:161— generateCustom button posts the custom conjecture text and shows proof buttons if confirmed
Node.js backend with standard frontendVerified
Proof Buster uses a standard frontend and a node.js based server for its backend
Claimed on Devposthigh confidenceserver.js:1— Express-based Node.js server serving static frontend and API endpointspackage.json:13— express dependency confirms Node.js backend
Random conjecture generator with filtersVerified
Implied by 'select a conjecture to prove' and gamified practice flow
Claimed on readmehigh confidencepublic/script.js:39— applyFilters filters conjectures by difficulty/tags checkboxespublic/script.js:103— generateConjecture click handler picks a random conjecture from the filtered set
Targeted feedback on proofsVerified
Offers targeted feedback to guide students to the correct answer
Claimed on Devposthigh confidenceserver.js:20— /give-feedback endpoint prompts Gemini to provide feedback and tips without giving the full solutionpublic/script.js:122— giveFeedback button calls /give-feedback and renders returned HTML feedback
Theorem/conjecture validity checkingVerified
It checks the validity of theorems
Claimed on Devposthigh confidenceserver.js:37— /check-proof endpoint sends conjecture and proof to Gemini and asks YES/NO on validitypublic/script.js:141— checkProof button calls /check-proof and displays Correct/Incorrect based on response
Gamified learning approachClaimed only
Provides students with an assisted gamified approach to proof writing inspired by platforms for programming and language learning
Claimed on Devpostmedium confidenceGemini fine-tuning/model tuning on conjectures and ready-made proofsClaimed only
The backend utilizes Gemini's API to train the model on conjectures and ready-made proofs to fine tune it to the project needs
Claimed on Devposthigh confidenceGemini function calling for non-String valuesClaimed only
Gemini's function calling to get non-String values for use throughout the project
Claimed on Devposthigh confidenceScores user proofsClaimed only
Scores user proofs
Claimed on Devpostmedium confidence
An AI agent derived these features from the project’s Devpost page and readme, then searched the code for each one. Verified features are backed by cited code; claimed-only features had no supporting code, which is not by itself proof a feature is missing.
Export this project's context (description, README, evidence, key source files) to chat with an AI agent elsewhere.