Generate speech locally using the pocket-tts CLI. Defaults to the `fantine` voice; use `george` when a second voice is needed.
Web search (text/news/images/videos/books) and URL content extraction using DuckDuckGo (ddgs and primp).
Transcribe audio/video files to text using OpenAI Whisper (faster-whisper backend). Supports word-level timestamps, multiple output formats (srt, vtt, json, txt, ass), auto language detection, and GPU acceleration. Use when transcribing audio, creating subtitles/captions, or extracting text from spoken content.
Extract clean text and structured Markdown from PDF files using PyMuPDF4LLM. Handles multi-column layouts, tables, headings, lists, and images. Use when converting PDFs to readable text, extracting content for LLM ingestion, or inspecting PDF structure. Does NOT handle scanned PDFs (no OCR).