Markets
The cost calculator tells you what you're paying. This tells you what you'd lose if you had to leave — data custody, application coupling, eval independence, exit readiness. Ten questions, no fence-sitting, straight to a bucket: Sovereign, Portable, Coupled, or Captive. Each one points somewhere concrete — Captive and Coupled stacks route to the Token-Cost Audit, Portable to Managed Inference, Sovereign to Stack Watch to hold the line you've earned.
01If your primary model provider doubled prices tomorrow, could you switch within a week?
02Does your product code call the provider SDK directly, or through a gateway you own?
03Do your prompts and system context live in your own repo, or only in the provider console?
04Do you have a written eval set for your top task, independent of any one provider?
05Do you know — for certain — whether your current tier trains on your data?
06Whose retention policy governs your prompt logs and caches — yours, or the provider’s?
07Have you actually run your production workload on a second provider — not just tested it?
08Do you have a zero-retention agreement or DPA in writing with your provider?
09Does the provider's forward-deployed or solutions team have standing access to your systems?
10Could you export every prompt and completion you have ever sent, today, if you needed to?
0 / 10 answered
0 / 20
Your stack is
Captive
Answer all 10 questions to see your bucket and next step.