Quick Start & Setup Guide

This guide walks you through compiling, running, and testing Project Anant on Linux.

---

Prerequisites

• Operating System: Linux (Ubuntu 22.04+, Debian 12+, Arch Linux, Fedora)
• C++ Compiler: GCC 13+ or Clang 17+ with C++26 support
• Build System: CMake 3.25+ and Ninja / GNU Make
• Libraries: DuckDB C library, OpenMP, liburing (for Aegon)
• Node / JS Runtime: Bun 1.1+ or Node.js 18+ (with npm)
• Hardware: Minimum 8 GB RAM (16 GB recommended), 4+ CPU cores

---

One-Click Launch

The simplest way to start the complete stack is via the root launcher script:

# Clone the repository
git clone https://github.com/Team-Zemo/ProjectAnant.git
cd ProjectAnant

# Start all services (builds C++ engine, builds React dashboard, starts Aegon on :3000)
chmod +x start.sh
./start.sh

Once running:

• Open your browser at http://localhost:3000 to access the interactive forensics studio.
• The engine automatically hosts the compiled React dashboard and REST/SSE API on port 3000.

---

Manual Step-by-Step Compilation

1. Build the Aegon Web Core & Anant Engine

cd anant-engine
cmake -B build \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_CXX_STANDARD=26

cmake --build build -j$(nproc)
cd ..

This compiles the high-performance binary at anant-engine/build/anant-engine.

2. Build the React 18 Dashboard

cd anant-dashboard
bun install       # Or: npm install
bun run build     # Or: npm run build
cd ..

The compiled bundle is output to anant-dashboard/dist/.

3. Run the Engine Daemon

./anant-engine/build/anant-engine --port 3000 --threads 8

---

Ingesting & Scoring 2,000,000 Transactions

Once the engine is running, trigger the complete SIMD ingestion, Bayesian scoring, and syndicate detection pipeline:

curl -X POST http://localhost:3000/api/ingest

Response Example

{
  "status": "ok",
  "message": "Ingestion and scoring started in background."
}

Checking Pipeline Status

curl http://localhost:3000/api/status

{
  "loaded": true,
  "ingest_running": false,
  "ingest_pct": 100,
  "score_pct": 100,
  "stage_num": 4,
  "stage": "Pipeline Complete",
  "message": "All records processed, mule scores computed, and graph ready for investigation.",
  "elapsed_ms": 3848,
  "duck_time_ms": 2718,
  "score_time_ms": 1054,
  "syn_time_ms": 76,
  "rows_loaded": 2000000,
  "unique_accounts": 24873,
  "victim_accounts": 300,
  "critical_mules": 1073,
  "syndicates_count": 122
}

The entire 2,000,000 transaction dataset is parsed, scored, and clustered in under 4 seconds.

---

Running Documentation Locally with VitePress

To run these documentation pages locally:

# In the project root:
bun run docs:dev    # Or: npm run docs:dev

Visit http://localhost:5173 to explore the docs with live hot reload.

To build the static HTML documentation site:

bun run docs:build  # Generates static site in docs/.vitepress/dist