ops
    New
    2026-04-13

    Evaluation Sandbox

    14-day hosted sandbox for product evaluation — pre-seeded with demo models, compliance evaluations, audit trail, and live LLM gateway.

    sandbox
    evaluation
    demo
    trial

    Overview

    The hosted evaluation sandbox at app.aegissovereign.io provisions a fully isolated workspace in under 30 seconds. Each sandbox includes 5 pre-configured demo models, pre-run compliance evaluations, 200+ audit events, data lineage DAG, bias evaluation results, a budget-capped LLM gateway, and pre-configured AI Governance Agents. The MCP Server is also available — connect Claude Code or Claude Desktop to your sandbox workspace using the API key shown on the sandbox dashboard.

    What's Pre-Seeded

    Every sandbox includes the same demo environment.

    ItemDetail
    FraudDetector v2Tabular classification — EU AI Act + DORA evaluated
    CreditRisk v3Credit scoring — SR 11-7 evaluated, intentional DIR bias finding
    ClinicalNLP v1Clinical text classification — HIPAA/FDA SaMD evaluated
    CustomerChurn v4Retention model — SOX/SEC evaluated
    AMLScreening v1AML transaction screening — MAS TRMG + NIST AI RMF evaluated
    Audit Trail200+ events across all models — verifiable hash chain
    Data LineageSnowflake + S3 sources with column-level PII tracking
    LLM GatewayConnected to a live model, budget-capped at $5 for evaluation
    AI Governance AgentsCompliancePipeline, IncidentResponse, RemediationSuggestion agents active
    MCP ServerConnect Claude Code/Desktop via the API key on the sandbox dashboard

    Sign Up

    Sign up at app.aegissovereign.io/sandbox/signup. Work email required — free providers (Gmail, Outlook, etc.) are rejected. Sandbox expires after 14 days and becomes read-only.

    bash
    1curl -X POST https://api.aegissovereign.io/api/v1/sandbox/signup \
    2  -H "Content-Type: application/json" \
    3  -d '{
    4    "email": "you@company.com",
    5    "full_name": "Jane Smith",
    6    "company": "Acme Financial",
    7    "role": "Head of MLOps",
    8    "primary_framework": "eu_ai_act"
    9  }'
    Edit this page on GitHub