Intelligent Asset Graph

Upload configs and logs – we'll infer your AI attack surface.

Drag-and-drop config files, logs, or small repo snapshots (.zip) from your AI stack. We parse endpoints, models, authentication methods, and dependencies to build an inferred asset graph. We never connect to your systems directly; everything is file-based and stays inside your tenant.

Start small: one or two services or a single environment (e.g., "customer-support chatbot" or "internal AI copilot") rather than full production dumps.

Supported: .zip .json .yml .log .txt .csv .pdf
Need a sample? Click to view a minimal test bundle you can recreate.

Create three files locally, put them into a folder, zip it as asset-graph-sample.zip, and upload that zip here.

1) ai-config.yml
services:
  - name: customer-support-bot
    type: llm_api
    model: gpt-4.1
    endpoint: https://api.openai.com/v1/chat/completions
    auth: bearer_token
    data_sources:
      - jira_tickets
      - zendesk_kb
    pii_exposure: medium

  - name: internal-code-copilot
    type: llm_api
    model: claude-3.5
    endpoint: https://api.anthropic.com/v1/messages
    auth: bearer_token
    data_sources:
      - github_enterprise
      - confluence
    pii_exposure: high
2) gateway.log
2025-11-20T03:15:22Z INFO  /v1/chat 200 user=cs-bot-prod model=gpt-4.1
2025-11-20T03:15:24Z WARN  /v1/chat 429 user=cs-bot-prod model=gpt-4.1 rate_limit_exceeded
2025-11-20T03:16:01Z INFO  /v1/agents 200 user=code-copilot model=claude-3.5
2025-11-20T03:16:05Z ERROR /v1/chat 500 user=unknown model=unknown unauthorized_api_key
3) models.json
{
  "models": [
    {"name": "gpt-4.1", "provider": "OpenAI", "environment": "prod"},
    {"name": "claude-3.5", "provider": "Anthropic", "environment": "prod"}
  ]
}

Zip these three files and upload the zip to see how the graph engine maps services, models, endpoints, and auth patterns.

© 2025 AI PQC Audit. Advanced multi-AI powered post-quantum cryptography security platform.

AI Quantum Dev
204 E. 2nd Ave. Suite 601
San Mateo, CA 94401

Powered by Proprietary Multi-AI Technology