AskGear Frontend

AskGear is a learning assistant focused on math and puzzle solving. Users can capture questions, let the recognizer extract structured text, and drive an agent-style conversation to analyse or explore related practice.

Setup

• yarn install
• cp .env.local.example .env.local and add VITE_OPENROUTER_API_KEY, DEFAULT_MODEL, MINI_MODEL
• yarn dev (Vite selects an available port automatically)
• yarn build / yarn test:recognize for verification

Core Features

• Quickly convert test papers or screenshots into readable question text, eliminating the tedious process of manual input.
• Track the AI's reasoning process in real-time using the "Analysis" tag to get a step-by-step explanation and the final answer.
• The "Similar Problems / Question Cards" tag allows learners to switch to related exercises or a card organization view at any time, connecting different knowledge points.

Further Guidance

Project architecture, import rules, and naming conventions are documented in AGENTS.md. Refer there before adding pages, features, or shared components. For code style and linting tips see the inline comments and package scripts above.