AXIOM-ONE Research Execution Engine

A Research-Grade Reasoning System (Not a Chatbot)

"Build a thinking-first research agent that decomposes problems, runs multi-step reasoning loops, validates its own outputs, and exposes the entire thought pipeline to the user."

This project is about epistemic rigor, not vibes.

---

Features

• Task Decomposition: Breaks complex questions into 4-7 atomic research tasks
• Subagent Execution: Each task runs in isolation with full reasoning traces
• Adversarial Validation: Hostile peer review catches contradictions and unsupported claims
• Manuscript Synthesis: Produces structured, inspectable research reports
• Real-time Progress: Live task tree updates with streaming progress
• Export Options: Copy to clipboard or download as Markdown (PDF coming in v2)
• Keyboard Shortcuts: Full accessibility support (Ctrl+Enter, ESC, Arrow keys)
• Cancel/Abort: Stop long-running executions gracefully

---

Architecture

The Pipeline: Decompose → Execute → Validate → Synthesize

sequenceDiagram
    participant U as User
    participant D as Decomposer
    participant E as Executor
    participant V as Validator
    participant S as Synthesizer

    U->>D: Research Question
    D-->>U: Task Graph (4-7 tasks)
    
    loop Parallel Execution
        U->>E: Execute Task
        E-->>U: Result + Reasoning
    end
    
    U->>V: Validate Results
    V-->>U: Flags & Confidence
    
    U->>S: Generate Report
    S-->>U: Final Manuscript

Loading

---

Installation

Prerequisites

• Node.js (v18+)
• Google Gemini API Key

Setup

1. Clone the repository:

   git clone https://github.com/yourusername/axiom-one.git
   cd axiom-one

2. Install dependencies:

   npm install

3. Configure environment:

   cp .env.example .env
   # Edit .env and add your Gemini API key

4. Run development server:

   npm run dev

---

Usage

1. Enter a research question (minimum 20 characters)
2. Click "Initialize Audit" or press Ctrl+Enter
3. Watch as the system decomposes your question into tasks
4. Monitor real-time progress as subagents execute
5. Review the final manuscript with confidence scores
6. Export results via Copy to Clipboard or Download MD

---

Tech Stack

• Frontend: React + TypeScript + Vite
• Styling: Tailwind CSS + Custom CSS Variables
• UI Components: shadcn/ui
• LLM: Google Gemini 2.0 Flash Thinking
• State: React hooks with streaming updates

---

Development

Build for production:

npm run build

Type checking:

npx tsc --noEmit

---

Why Axiom-One Matters

Most "AI agents" optimize for fluency, hide uncertainty, and collapse under scrutiny.

Axiom-One optimizes for correctness under pressure, makes uncertainty explicit, and treats reasoning as a first-class artifact.

---

License

MIT