Model Modes

Model modes define the primary function of a model in no_llm. Each mode represents a specific type of task the model is designed to perform.

Limited Mode Support

Currently, no_llm primarily supports CHAT mode and chat-based models. Support for other modes (completion, embedding, image generation, audio) is under development and will be available in future releases.

Available Modes

Mode	Description	Common Use Cases
CHAT	Interactive chat completion	- Conversational AI - Virtual assistants - Interactive Q&A
COMPLETION	Text completion	- Text generation - Content creation - Code completion
EMBEDDING	Vector embedding generation	- Semantic search - Document similarity - Text clustering
IMAGE_GENERATION	Image creation from text	- Art generation - Design mockups - Visual content creation
AUDIO_TRANSCRIPTION	Speech-to-text conversion	- Meeting transcription - Subtitle generation - Voice notes to text
AUDIO_SPEECH	Text-to-speech synthesis	- Audio content creation - Accessibility features - Voice assistants

Listing Models by Mode

You can list all models supporting a specific mode using the registry:

from no_llm.config.enums import ModelMode
from no_llm.registry import ModelRegistry

registry = ModelRegistry()

# List all chat models
chat_models = list(registry.list_models(mode=ModelMode.CHAT))

# List chat models from a specific provider
openai_chat_models = list(registry.list_models(
    mode=ModelMode.CHAT,
    provider="openai"
))