Deployed Model

Interact with ML models deployed to remote endpoints

For many models, running them in the browser won't suffice. For these applications, we've built the Deployed Model class. To use this class you point to a model endpoint and provide a key, and this enables you to automatically pass that endpoint any information that was gathered at the harvesting step and transformed at the preprocessing step.

This class powers the majority of our LLM-based use cases, including our community version of the AI Squared Chrome extension, which makes use of OpenAI's GPT 3.5 API.