An AI-powered academic assistant built with React, TypeScript, and Google Gemini. It helps users truly understand complex content through logical breakdown, heuristic questioning, and Feynman testing.
- 🧠 Logic Decomposition: Automatically identifies core arguments and evidentiary structures in complex texts.
- ❓ Heuristic Questioning: Generates deep-thinking questions using Chain-of-Thought (CoT) to stimulate critical analysis.
- 🧩 Knowledge Synthesis: Summarizes content into modular, digestible insights.
- 🎓 Feynman Test Generation: Creates targeted assessment quizzes to verify true knowledge internalization.
- 🎨 Modern UI: Beautifully crafted interface with smooth animations and responsive design.
The DeepInsight Learning Agent follows a structured multi-stage reasoning process to ensure deep internalization of knowledge.
graph TD
A[Raw Text Input] --> B{AI Context Analysis}
B --> C[Logic Decomposition]
B --> D[Heuristic Questioning]
B --> E[Feynman Quiz Generation]
C --> F[Logic Map Visualization]
D --> G[Deep Thinking Prompts]
E --> H[Self-Assessment Quiz]
F & G & H --> I[Knowledge Internalization]
When running the ingestion process, the agent provides detailed feedback on its reasoning steps:
🚀 [AGENT START] Initiating Deep-Learning Workflow...
📊 [INPUT] Document received. Length: 5420 characters.
🔍 [PROCESS] Step 1: Analyzing document structure and semantic clusters...
↳ [Thought] Scanning for key concepts and logical transitions...
✅ [SUCCESS] Semantic structure parsed.
🧠 [PROCESS] Step 2: Generating high-order reasoning questions...
↳ [Thought] Identifying logical gaps in the text...
✅ [SUCCESS] 2 deep-thinking questions generated.
🌐 [PROCESS] Step 3: Executing web-based knowledge retrieval...
↳ [Tool Call] Searching: 'How does non-locality challenge the principle of local realism?'
↳ [Status] Connecting to DuckDuckGo Search API...
✅ [SUCCESS] External knowledge integrated into context.
📝 [PROCESS] Step 4: Synthesizing Feynman-style learning package...
↳ [Thought] Mapping extracted knowledge to pedagogical patterns...
✅ [SUCCESS] Learning package finalized.
🌈 [WORKFLOW COMPLETE] All tasks finished successfully.
- Frontend: React 19, TypeScript, Vite
- Styling: Tailwind CSS
- Animation: Motion (formerly Framer Motion)
- AI Engine: Google Gemini API
- Icons: Lucide React
- Node.js 18+
- A Google Gemini API Key
# Clone the repository
git clone https://github.com/yourusername/deepinsight-learning-agent.git
# Navigate to project directory
cd deepinsight-learning-agent
# Install dependencies
npm installCreate a .env file in the root directory (never commit this to GitHub):
VITE_GEMINI_API_KEY=your_gemini_api_key_herenpm run devDistributed under the Apache-2.0 License. See LICENSE for more information.
Built with ❤️ for lifelong learners.