3.11.1

Back to 41 More Apps

• Dictate from the keyboard and VivaDicta returns you to 41 more apps than before
• Another 24 apps that publish no way back are now recognized as such, so instead of trying and failing they take you straight to the manual return with your recording already running

One Model per Provider

• The model list on the AI Providers screen used to show the ten highest-scoring models, and four of those slots went to other models from a provider already on the list
• It now shows the best model from each provider, which answers the question that screen is for: which provider to set up. Each provider's other models are still in its own picker
• Five of the recommended models come with a free tier, and they are marked as such
• Mistral joins the list with mistral-medium-latest, which scores 9.4 for cleanup quality and answers in under half a second on a free tier

Fixes and Refinements

• When Grok is not set up yet, the mode editor now offers Grok sign-in directly instead of the plain API key screen, since both Grok speech-to-text and Grok text processing accept the Grok subscription
• The link under the model list now states the right number of measured models