You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Added model-list discovery to the local Settings page for both LLM and Embedding configuration.
Added a searchable fetched-model picker while keeping the model field manually editable.
Added provider discovery support for Ollama, OpenAI-compatible endpoints, OpenAI, Anthropic, Google, and Custom providers; Azure returns an unsupported-provider message in this first version.
Added POST /api/config/models so the server fetches model lists with saved or newly entered provider credentials instead of exposing third-party API keys to the browser.
Marked cloud mode as unsupported for model discovery and surfaced the limitation in Settings.
Added local browser-origin checks for model discovery proxying, plus structured errors for missing credentials, endpoint failures, auth failures, timeouts, and parse failures.
Added regression coverage for model discovery services, routes, client API normalization, Settings wiring, and i18n copy.