Verba

Verba is an open source application designed to provide an end-to-end, smooth and user-friendly Retrieval Augmentation Generation (RAG) interface. It combines state-of-the-art RAG technology with Weaviate's context-aware database and can be deployed on-premises or via the cloud to easily explore data sets and extract insights.

The target audience is developers, data scientists and enterprise users who need a powerful tool to process and analyze large amounts of data. Verba provides highly customized and automated data exploration solutions to help them obtain information and insights more effectively.

Data scientists use Verba to analyze research data and extract key insights. Enterprise users use Verba to optimize customer service and provide instant assistance through chatbots. Developers integrate Verba into their applications to enhance data query and processing capabilities.

Product features include:

Supports local and cloud deployment, providing flexible query and interaction methods.

Combined with Weaviate's context-aware database, efficient data retrieval and generation is achieved.

Supports multiple RAG frameworks, data types, chunking and retrieval techniques, as well as different large language model (LLM) providers such as OpenAI, Cohere, Google, etc.

Provides the function of importing multi-format data such as PDF, CSV/XLSX.

Supports the combination of semantic search and keyword search to improve search accuracy and efficiency.

It has semantic caching function to save and retrieve results according to semantic meaning.

Provide automatic completion suggestions to improve user experience.

Usage tutorial:

1. Install the necessary Python environment and dependencies.

2. Install Verba via pip or build from source.

3. Configure the required environment variables, such as Weaviate URL, API key, etc.

4. Start the Verba application.

5. Use the management console to import the data.

6. Ask relevant questions on the chat page and obtain semantically relevant data chunks and generated answers.

7. Configure the RAG pipeline as needed.