Overview
Verba is a community-driven open-source application designed to offer an end-to-end, streamlined interface for Retrieval-Augmented Generation (RAG). It allows users to explore datasets and extract insights either locally with Ollama and Huggingface or through LLM providers like Anthropic, Cohere, and OpenAI. Verba combines RAG techniques with Weaviate's context-aware database, enabling users to query and interact with data. It supports different RAG frameworks, data types, chunking & retrieving techniques, and LLM providers. Verba can be deployed via Docker or built from source, making it accessible for various deployment environments. It supports importing data from various sources including UnstructuredIO, Firecrawl, UpstageDocumentParse, PDF, GitHub, GitLab, and CSV/XLSX files.