- Only delete LM Studio models on a local server with a single variant
- Count every loaded LM Studio copy in memory totals
- Cap the saved context window at the model's limit on Use with Cline
- Clear the runtime model list cache after downloads and deletes
- Hide the local model chip in SSH sessions
- Ignore stale status refreshes and stale ClinePass lookups
- Add fetchCurrentUserPlan to RpcClineAccountService
- Label the Ollama response timeout as the default
Saoud Rizwan
aa52079cc7