A reasoning-first AI workspace built to eliminate conversation drift and coherence loss in long AI chats. Crucible anchors vague ideas in a persistent, traceable execution plan, visualising them as a structured node graph alongside your chat.
Designed for solo creators, student builders, and hackathon competitors developing complex concepts under pressure for the USAII Global AI Hackathon 2026.
When you brainstorm with traditional AI chatbots, the reasoning is invisible. You get a long chat transcript, but no shared structure of what was decided, what was rejected, or why.
Crucible solves this by externalising the AI's reasoning in real time onto an interactive Bubble Canvas. It actively interrogates your concept, checks for contradictions, and builds a concrete execution plan where every milestone traces back to the specific assumption it validates.
- Spotlight Capture: Centered, minimal frosted-glass input bar to jump-start your idea development.
- The Bubble Canvas: Interactive, organic React Flow visual node graph mapping core claims, target users, assumptions, and scope boundaries.
- Contradiction Engine: Real-time semantic checking of new requests against existing assumptions, flagging logical tensions with consequence links.
- Traceable Roadmap: Transition the canvas into a step-by-step execution plan where every milestone traces back to the specific assumption it validates.
- The Verdict: Non-automated, human-in-the-loop Go/No-Go options highlighting the riskiest assumptions.
- Frontend: React 19, TypeScript, Vite 8, React Flow, Framer Motion, Tailwind CSS v4
- Backend: FastAPI, Uvicorn, SQLite, Pydantic v2
- AI Integration: Google Gemini API (
gemini-2.5-flash), Gemini Search Grounding
- Python 3.10+
- Node.js 18+
-
Clone the repository and open the project
git clone https://github.com/adrienoke/Crucible.git cd Crucible -
Start the Backend
cd backend # Create virtual environment python -m venv .venv # Activate virtual environment # On Windows: .venv\Scripts\activate # On macOS/Linux: source .venv/bin/activate # Install dependencies pip install -r requirements.txt # Create a .env file and set your Gemini API key: # GEMINI_API_KEY=your_api_key_here # Start the FastAPI server uvicorn main:app --reload
The backend API runs locally on
http://127.0.0.1:8000. -
Start the Frontend In a new terminal window:
cd frontend npm install npm run devThe frontend dev server runs locally on
http://localhost:5173.
To run backend test suites:
cd backend
pytestContributions are welcome! Please follow these basic steps:
- Fork the repository.
- Create a new feature branch (
git checkout -b feature/amazing-feature). - Commit your changes (
git commit -m 'Add amazing feature'). - Push to the branch (
git push origin feature/amazing-feature). - Open a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.