Offline Local Mode
No cloud AI provider is contacted for the supported local workflow.
Local mode
Use supported local workflows without contacting a cloud AI provider, and move to approved cloud processing only when the workflow and user action require it.
Three processing modes
Privacy behavior is specific to the enabled workflow. Chameleon does not describe every feature as local or every request as cloud-based.
No cloud AI provider is contacted for the supported local workflow.
Only user-approved content is sent to approved cloud AI infrastructure.
The approved request payload is processed through server-side provider routing.
Desktop app and local gateway
Availability depends on the installed client, the enabled feature, and the specific workflow.
Use short-lived or scoped credentials rather than embedding master provider keys in the client.
Keep supported file handling within the local workflow until the user approves a cloud step.
Session and device access are handled separately from provider credentials.
Revoke supported device access without exposing a provider master key.
Enterprise private deployment, data-residency controls, zero-data-retention arrangements, role permissions, and audit logging depend on the verified implementation and contract scope. Planned controls are not represented as completed features.