What's New
Ollama Model Picker (#221)
When using Ollama as your LLM or embedding provider, model name fields in Settings now include a dropdown button that fetches available models from your Ollama instance via /api/tags. You can still type a model name manually for full flexibility. The list refreshes automatically when the Ollama URL changes.
Enhanced Download Progress (#222)
The model download progress bar now shows additional detail during downloads:
- Total size and amount downloaded (e.g., "142 MB / 800 MB")
- Current download speed (e.g., "3.5 MB/s")
- Estimated time remaining (e.g., "2m 15s remaining")
Size and speed are shown for models with known download sizes. ETA is computed for all models from the progress rate.
Contributors
Thanks to @kerolloz for both feature suggestions!