Multimodality & Senses

Vision Support (Multimodal)

• Image Integration: Drag-and-drop or upload images for analysis (requires a multimodal-capable GGUF like Llava or Moondream).
• Base64 Processing: Local, secure image encoding.

File Contextualization

• Document Indexing: Attach code files (.go, .js, .py) or documents (.md, .txt) to give the AI massive instant context for a specific task.

Local STT (Speech-to-Text)

• Offline Transcription: Record voice messages directly in the app.
• Bundled Engine: Uses a localized environment (Vosk/Whisper equivalent) for zero-latency, private transcription.

Source document