Step 1 — Provision a Sandbox

Go to app.aegissovereign.io and sign up with your work email. Your sandbox is provisioned in under 30 seconds with 5 pre-configured demo models, compliance evaluations already run, 200+ audit events, and a live LLM gateway. No credit card required.

Sandbox workspace — Your provisioned sandbox — pre-seeded models, KPIs, and system health, ready to explore immediately.

Step 2 — Register Your First Model

Navigate to Model Registry → Register Model. Fill in the model name, version, framework (scikit-learn, PyTorch, TensorFlow, etc.), task type, and training dataset reference. The platform assigns status registered and triggers the CompliancePipelineAgent automatically.

bash

1curl -X POST https://sovereign.yourcompany.com/api/v1/registry/models \
2  -H "Authorization: Bearer $API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{
5    "name": "fraud-detector-v1",
6    "version": "1.0.0",
7    "framework": "scikit-learn",
8    "workspace_id": "ws-prod",
9    "task_type": "classification",
10    "metadata": { "training_dataset": "transactions-2024-q4", "accuracy": 0.943 }
11  }'
12# → { "id": "model-7f3a", "status": "registered" }

Step 3 — Run a Compliance Evaluation

Navigate to Compliance → Run Evaluation. Select your model and the regulatory framework for your jurisdiction (EU AI Act for EU, SR 11-7 for US banks, MAS TRMG for Singapore). The evaluation completes in under 5 minutes and produces a scored report with findings and a signed compliance artifact.

typescript

TypeScript SDK

1import { AegisSovereignClient } from "@aegissovereign/sdk";
2
3const client = new AegisSovereignClient({
4  baseUrl: "https://sovereign.yourcompany.com",
5  apiKey: process.env.AEGISSOVEREIGN_API_KEY!,
6  workspaceId: "ws-prod",
7});
8
9const result = await client.regulatory.evaluate("fraud-detector-v1", "eu_ai_act");
10console.log(result.passed);    // true / false
11console.log(result.score);     // 0.83
12console.log(result.findings);  // [{control, severity, recommendation}]

Step 4 — Verify the Audit Chain

Every action — model registration, evaluation, promotion — produces an audit entry cryptographically chained to the previous one. Navigate to Audit Log → Verify Chain to confirm chain integrity. The chain can also be verified via the API at any time for regulatory submissions.

bash

Verify chain integrity

curl -X POST https://sovereign.yourcompany.com/api/v1/audit/verify/latest \
  -H "Authorization: Bearer $API_KEY"
# → { "chain_valid": true, "entries_verified": 247 }

Next Steps

Now that your first model is registered and evaluated, explore the rest of the platform.

Feature	Where to find it	Doc
Champion/challenger A/B split	Model Registry → A/B Split	`model-registry`
LLM Gateway + data classification	LLM Gateway → Route Simulator	`llm-gateway`
Real-time drift monitoring	Monitoring → Drift Dashboard	`drift-detection`
GitOps policy bundles	Policy Center → Bundles	`gitops-policy`
MCP Server for Claude Code	Settings → MCP Server	`mcp-server`