Whispering, Buzz and LM Studio: everyday AI apps

In this Portuguese-language video, the speaker reviews ten apps used in their workflow, with AI tools first and conventional utilities afterward. This is an overview of uses and features rather than a setup walkthrough.

LM Studio provides a graphical way to run models locally. The speaker describes model downloads, RAM-based quantization guidance and a local server compatible with the OpenAI API, comparing it with Ollama. Buzz transcribes audio or video files with Whisper and can export timestamped subtitles. Whispering uses a global keyboard shortcut to turn speech into text at the cursor. The speaker uses it to dictate longer coding prompts, but does not demonstrate its transcription-provider configuration. The presenter describes both transcription apps as open source and available for Windows, Mac and Linux.

Google Opal builds AI workflows from written instructions in a browser-based visual canvas. Its inclusion does not establish that it runs locally. Google AI Edge Gallery is a separate phone-focused option: the speaker describes downloading Gemma 4 E2B or E4B for offline use on Android or iPhone, though the quality claim concerns more powerful phones.

The remaining apps cover routine computer tasks. Espanso expands text shortcuts and supports dynamic values; ActivityWatch records activity locally for later review. Transmute is a self-hosted file converter with an API. Stirling-PDF runs in Docker and handles PDF operations such as merging, compression and OCR. iTerm2 adds split panes, output search and profiles to a Mac terminal workflow.