The Windows version of the Giga Pisar dictation tool has received an update to version 1.0.5, introducing a feature called 'Brain' that allows users to connect external AI models. Users can integrate services like DeepSeek, OpenRouter, OpenAI, Groq, Gemini, or Claude via an API key, or use local servers like Ollama and LM Studio.
The integration process has been simplified: the application automatically detects the service provider based on the API key format and selects the appropriate model. Security measures include encrypted key storage and a system tray indicator that shows the status of the AI connection. By default, the feature is disabled, and only text is sent to the server.
The update also adds a new keyboard shortcut, Left Ctrl + Win, for dictation. Users can now highlight text in applications like Word or browsers and issue voice commands such as 'shorten' or 'translate'. A 'Correct every dictation' option has been added to automatically process all voice input through the selected AI model.
The macOS version of Giga Pisar has been updated to version 3.7.1, incorporating similar cloud-based AI functionality. The developer has unified the settings interface across both platforms, moving to a section-based layout similar to Windows 11 settings. Future development plans include implementing an offline 'Brain' for Windows.
