Documentation

Understand local and cloud inference

Follow the data, not just the product name.

Step 1

Identify where the selected model runs: your computer, a cloud provider or rented hardware. Remote execution sends the request and relevant context away from your device.

Step 2

Separate model inference from account information, conversation storage and platform metadata. A local model does not prove that messaging or account data remain local.

Step 3

Review the applicable data policy before sharing sensitive content. Conversation history stores messages; assistant memory retains selected information; training updates model parameters. None implies the others automatically.

Open GecoChat