Application Templates FAQs

What is Cognita? How can I use it with and without TrueFoundry?

Congnita is our open-source RAG framework. Cognita uses Langchain/Llamaindex under the hood and provides an organisation to your codebase, where each of the RAG components is modular, API-driven and easily extendible.

While Cognita can be used independently, utilizing it on TrueFoundry can significantly enhance your end-to-end LLM workflows. TrueFoundry provides optimized infrastructure and implements the best SRE principles with the right governance for LLM deployment, Gateway, Finetuning etc.

What vectorDBs do you support in RAG?

We currently support Qdrant and Singlestore. However, our modular framework allows you to implement any vector database by extending the vectordb base class.

I can't find the parser/embedder that I want to use in the RAG framework.

All the components of the RAG framework are customizable. You can extend the base class of DataLoader, Embedder, Parser and VectorDBs and write your own implementation.

Please refer to this for detailed steps.

How can I optimize indexing with large datasets which might have duplicate entries?

We support incremental indexing that ingests entire documents in batches (reduces compute burden), keeps track of already indexed documents and prevents re-indexing of those docs.