ARGUS — AI Usage Intelligence Platform

Governing AI adoption — before liability hits.

---

🛡️ The Mission

Your employees are using ChatGPT, Claude, and Gemini every day. They're pasting source code, customer data, and internal documents. That data is leaving your company right now, without visibility and without control.

This isn't a hypothetical. This is reality. Leaked secrets, hallucinated packages entering your codebase, and spiraling costs. ARGUS makes all of this visible. We help companies govern AI adoption before it becomes a liability.

---

🚀 Key Capabilities

Feature	Description
Critical Findings	Real-time detection of leaked AWS keys, PII (emails, IPs, IBANs), and database credentials.
Slopsquatting Scan	Detect hallucinated or malicious packages (e.g., reqests, pandass) suggested by AI.
Provider-Agnostic	One dashboard for all tools. Monitor OpenAI, Anthropic, Gemini, and local Ollama models.
Compliance Score	Automated audit trails and risk documentation ready for the EU AI Act.
Cost Intelligence	Aggregate spend by department and model with recommendations for 40%+ cost savings.

---

📦 The "Ready-to-Go" Lab Device

We focus on data integrity. Once your data leaves your network, you've lost control. We ship ARGUS as a ready-to-go lab device (or on-prem software) that proves your risk exposure.

• Deterministic & Agentic: Combines high-speed regex scanning with LLM-powered sensitivity analysis.
• Edge Deployment: Runs entirely inside your walls on affordable hardware like a Raspberry Pi.

---

🛠️ Architecture

flowchart TB
    subgraph "Data Sources (Live & Batch)"
        PROXY[Live Proxy / API]
        CHAT[Chat Exports<br/>ChatGPT, Claude, Gemini]
        JSONL[JSONL Logs<br/>IDE / CLI logs]
    end

    subgraph "Analysis Pipeline"
        WORKER[Analysis Worker]
        
        subgraph "Deterministic Analyzers"
            SECRETS[Secrets Scanner<br/>API keys, tokens]
            PII[PII Detector<br/>Emails, IPs, IBANs]
            SLOP[Slopsquatting Detector<br/>Hallucinated packages]
        end
        
        subgraph "Agentic Analyzers (LLM)"
            SENSITIVITY[Sensitivity Analysis]
            QUALITY[Usage Quality Scan]
            COMPLIANCE[EU AI Act Scoring]
        end
    end

    subgraph "Storage & Intelligence"
        PG[(PostgreSQL<br/>Audit Trails)]
        VIEWS[Materialized Views<br/>Risk Aggregates]
        REC[Recommender Agent<br/>Cost/Security Advice]
    end

    subgraph "Command Center"
        FASTAPI[FastAPI Backend]
        REACT[React Dashboard]
    end

    PROXY --> WORKER
    CHAT --> WORKER
    JSONL --> WORKER
    
    WORKER --> SECRETS & PII & SLOP
    WORKER --> SENSITIVITY & QUALITY & COMPLIANCE
    
    SECRETS & PII & SLOP --> PG
    SENSITIVITY & QUALITY & COMPLIANCE --> PG
    
    PG --> VIEWS
    VIEWS --> REC
    REC --> VIEWS
    
    VIEWS --> FASTAPI
    FASTAPI --> REACT

Loading

---

🏁 Quickstart for Judges

Follow these steps to launch the ARGUS production environment.

1. Ensure Docker is installed and the domain/TLS is configured (for production).
2. Setup Environment:

   cp .env.example .env
   # Edit .env and populate GOOGLE_API_KEY

3. Start Platform:

   docker compose up -d

Access URLs:

• Frontend Command Center: http://localhost:3000

---