llm-app

llm-app is a repository of ready-to-deploy LLM app templates built on the Pathway Live Data Framework's AI Pipelines. According to the project, these pipelines let developers put AI applications into production that offer high-accuracy RAG and AI enterprise search at scale using the most up-to-date knowledge available in their data sources. The templates can be tested on a local machine and deployed on-cloud (GCP, AWS, Azure, Render) or on-premises.

The apps connect and sync with data sources on the file system, Google Drive, Sharepoint, S3, Kafka, PostgreSQL and real-time data APIs, covering new data additions, deletions and updates. They include built-in data indexing enabling vector search, hybrid search and full-text search, done in-memory with cache, and come with no infrastructure dependencies requiring separate setup.

The repository lists templates including a Question-Answering RAG App, Live Document Indexing (vector store / retriever), a Multimodal RAG pipeline with GPT-4o, an Unstructured-to-SQL pipeline with SQL question-answering, Adaptive RAG, a Private RAG App with Mistral and Ollama, Slides AI Search, and Video RAG with TwelveLabs. The templates are described as scaling up to millions of pages of documents, and pipeline steps can be changed, such as adding a new data source or switching a vector index to a hybrid index.

The apps can be run as Docker containers and expose an HTTP API to connect a frontend, and some templates also include an optional Streamlit UI. They rely on the Pathway Live Data Framework for data source synchronization and for serving API requests.