Models

Pick which models power text (planning, captions), images, and video. Add Replicate models by pasting from the website — no code edits.

Text models

Used for planning, captions, and video briefs. Add a Replicate LLM, or use local Ollama if you prefer.

ActiveLabelSourceActions
Ollama (local) config.jsonollama_url / ollama_model Requires ollama serve
llama-3.2-3b justmalhar/meta-llama-3.2-3b {} Edit
gpt-5-mini openai/gpt-5-mini {} Edit
yes meta/meta-llama-3-8b-instruct meta/meta-llama-3-8b-instruct {"max_tokens": 2048, "temperature": 0.7} Edit

+ Add text model (Replicate)

Image models

+ Add image model

ActiveLabelSlugDefaultsActions
yes openai/gpt-image-2 openai/gpt-image-2 {"aspect_ratio": "1:1", "quality": "low", "output_format": "png"} Edit defaults

Video models

+ Add video model

No video models yet. Video slots will only produce a brief (no clip) until you add one.

Back to brands