Kuudo has one proposition and two deployment choices. Managed MCP runs the services for you. Private Cloud runs them in your cloud account. Both preserve your choice of AI client, your Amazon accounts, your knowledge and your operating rules. Neither path is a required upgrade from the other.
Choose the infrastructure arrangement that fits
| Choice | What you choose it for | Where to start |
|---|---|---|
| Managed MCP | Connect without deploying the services | Managed setup |
| Private Cloud | Control the service infrastructure in your cloud | Private Cloud |
Ads, Selling Partner, Vendor Central, Atlas and Iris are available through Managed MCP. Agent Flow's pipelines and customer-owned data lake remain Private Cloud. Choosing Managed MCP does not create an Agent Flow lake or automatically retain account history.
Inventory the work before changing deployment
You can move between Managed MCP and Private Cloud. Plan the change around the connections and artifacts your team actually uses:
- The AI clients, automations and service connections to switch.
- Amazon account grants, advertising profiles, marketplaces and access roles.
- Your installed Skills, custom instructions, operating rules and approval policies.
- Generated assets, their storage locations, download access and any service state needed to continue work.
- For Agent Flow, the pipelines and lake in your Private Cloud deployment.
Changing the endpoint alone does not transfer account authorization, assets or retained service state. Confirm the transition steps with the Kuudo team before changing a production workflow.
Verify the new connection before switching the work
- Obtain the new deployment's connection details and configure the required service credentials.
- Resolve the intended Amazon accounts and permissions. Confirm any reauthorization needed for the new runtime.
- Connect an AI client and run a small read against the expected account, or a source lookup for Atlas.
- Restore your Skills and operating rules, then check the approvals for any proposed write or billed generation.
- Switch workflows after the new path is verified. Revoke superseded credentials and close the old service arrangement when the transition is complete.
Keep asset access and active jobs in the transition plan. An Iris video edit or extension can depend on retained generation lineage; preserve the required state rather than treating a downloaded file as a complete migration.
Keep the same ownership principles
The deployment changes who operates the infrastructure. Your choice of AI client, your business knowledge, your account decisions and your rules remain central to the work. Read the service's current data-handling and commercial terms when choosing either path.
Discuss a deployment move with the Kuudo team, or return to the Managed MCP overview.