Model Catalog

Default category mapping:

• textToText: deepseek-ai/deepseek-v3
• textToSpeech: minimax/speech-02-turbo
• textToImage: prunaai/z-image-turbo
• imageTextToVideo: wan-video/wan-2.2-i2v-fast

Alternative model metadata is stored in models/ and validated against schema rules before execution.

See:

• Configuration
• Model Adapters and Options
• Glossary