Introducing custom text encoders on Vectara

Vectara now allows users to create and manage custom text encoders, offering seamless integration and enabling tailored text embedding processes that align with specific application needs. This feature provides flexibility by allowing the choice of suitable text embedding models, whether hosted on-premises or through external providers, and ensures consistency by standardizing text processing across applications. Users gain control over encoder management, with the ability to update encoders as project requirements evolve, while maintaining security through the use of vetted models. The integration of an OpenAI API-compatible encoder is facilitated through a POST request to the Create an encoder API, where users can specify various parameters such as type, name, and model. This expanded encoder management capability offers greater control and flexibility for text processing within the Vectara platform, and users are encouraged to engage with the community for feedback and further exploration of Vectara's offerings.