Model slots with fallbacks
Assign a primary and a backup model to each capability. If a model refuses or a provider errors, the request falls through to the next one instead of failing the conversation.
AI provider configuration
Connect an AI provider once, then choose models for chat, embeddings, reranking, speech, and transcription — each with a backup — from hundreds of free and paid options.
See the application
Assign a primary and a backup model to each capability. If a model refuses or a provider errors, the request falls through to the next one instead of failing the conversation.
Models are listed with their pricing so you can run entirely on free models while evaluating, then move to paid models without changing anything else.
Provider keys are encrypted with AES-256-GCM before storage and are never shown again after saving — the interface displays only the last four characters.
IN PRACTICE
In Dino Inventory, yes. Models are selected per capability — chat, embeddings, reranking, speech, and transcription — from the provider you connect, and each slot takes a backup model. Free and paid models appear together with their pricing, so a workspace can run on free models and switch later.
A clearer way to work
Purchasing, warehouse work, and fulfillment. Connected.