Llm App

LLM App is a set of ready to deploy application templates built on the Pathway Live Data Framework, a data processing framework with a streaming engine written in Rust. The templates cover retrieval augmented generation and ETL pipelines, and are meant to put AI applications into production that answer from the most current data in connected sources such as local files, Google Drive and SharePoint. Each template is defined with YAML and Python and can be customized by developers and non developers alike. The RAG templates include a basic question answering app that uses a GPT model of choice over documents such as PDF and DOCX files, an adaptive RAG app that the vendor says reduces token cost by up to 4 times while keeping accuracy, and a fully private local version using Mistral and Ollama. There is also a multimodal pipeline that uses GPT 4o when parsing documents with charts and tables, and a live document indexing service that works as a vector store. Pipelines update as documents change or new ones arrive. Pathway publishes pricing for the framework itself. The Community edition is free and open under the BSL 1.1 license, self hosted only, and supports up to 8 GB of RAM on 4 cores. The Scale tier supports up to 16 GB of RAM on 4 cores and is available free or paid with a license key. Enterprise licensing is available for self hosted and managed use, and supports up to 24 TB of RAM across 40 nodes. No amounts are listed for the paid tiers.