Explore a different way to prepare coding practice. Whetstone pairs generated Python exercises with reference solutions and tests, then records which candidates pass or fail in a reviewable log. Available as a source project.
Created by Weisley Dan Ouma
Built with Codex · GPT-5.6
Was this build worth discovering? Vote without signing in.
The cost of using this project has not been verified. Check the creator’s app for current terms.
Source checked September 22, 2026Creator Devpost submission and linked repository explicitly describe building the core exercise generation and checking workflow with Codex and GPT-5.6.
Start with similar ideas, then find something unexpected.
App previewStuck on one step in a mathematical paper? Interactive Proof lets learners select a passage in an uploaded PDF and request an explanation beside the source, with optional Lean code for context.
▶ Demo previewExplore anatomy as sculpture. A marble-style scene turns muscle groups into selectable parts.
▶ Demo previewLook inside a car in 3D. Explore its parts and follow lessons that bring the mechanics into view.
App previewRestore music to a silent world. BEET combines side-view combat with an isometric garden, where magical plants contribute musical layers as you explore. The creator-linked Unity demo is intended for desktop browsers.
App previewGive each plant its own history. GardenCircle AI combines photo identification, editable plant profiles and weekly photo check-ins with care questions and plant-exchange listings. This open-source prototype requires local setup.
App previewDescribe a terminal task in everyday language. CLAI proposes a shell command with an explanation and risk label, then lets you review, edit or reject it. It supports several AI providers and local Ollama models.