Media resources

Two resources that pull text out of a media file. Each has its own reference page. OCR and offline transcription need no API key and no network.

Works in workflow and agent mode.

Resource	Extracts text from	Reference
component: transcribe_audio	Audio (speech to text via a Whisper API)	Transcribe
component: transcribe_audio_offline	Audio (speech to text offline via whisper.cpp)	Transcribe
component: ocr_image	Images (printed or handwritten text via tesseract)	OCR

See also

• LLM resource - vision models can read images directly in a prompt
• Image analysis tutorial - files: on a chat prompt instead of OCR
• RAG resources - index the extracted text for search
• Resources overview - all resource types