Model picker in the Web App

Goal

Choose which model handles a conversation and where it runs — cloud, browser or your device.

What you need

• Access to the Web App at app.botconnector.id.
• For device runs: the Local Runtime installed (Install the Local Runtime).

Steps

1. Open a chat in the Web App.
2. Open the model picker in the conversation header.
3. Choose a model. The picker shows what the model can do (chat, tools, vision, coding — see Capabilities) and which compute targets it can run on.
4. Choose the compute target: Cloud runs through BotConnector Cloud (Use Cloud AI); Local runs on your machine through local Ollama (Run a model). Browser-local inference is planned (Run a model in your browser).
5. Send a message. The conversation shows which target handled it.

Expected result

Your conversation runs on the model and compute target you picked, and the target stays visible in the chat.

If it fails

• A model does not appear or will not respond: check the status page for the capability's real state, then Troubleshooting.
• Cloud model unavailable: availability depends on current account access and service capacity — see Cloud.
• Local model will not load: Memory problems or Runtime problems.

Next step

Attach files to the conversation: Files, or search the web from a chat: Web search.