multimodal

Examples for image/audio/video processing patterns.

Files

• audio_input_output.py - Demonstrates audio input output.
• audio_sentiment_analysis.py - Demonstrates audio sentiment analysis.
• audio_streaming.py - Demonstrates audio streaming.
• audio_to_text.py - Demonstrates audio to text.
• image_to_audio.py - Demonstrates image to audio.
• image_to_image.py - Demonstrates image to image.
• image_to_structured_output.py - Demonstrates image to structured output.
• image_to_text.py - Demonstrates image to text.
• media_input_for_tool.py - Demonstrates media input for tool.
• video_caption.py - Demonstrates video caption.

Prerequisites

• Load environment variables with direnv allow (including OPENAI_API_KEY).
• Create the demo environment with ./scripts/demo_setup.sh, then run cookbooks with .venvs/demo/bin/python.
• Some examples require optional local services (for example pgvector) or provider-specific API keys.

Run

• .venvs/demo/bin/python cookbook/02_agents/<directory>/<file>.py