Verified project record

pathwaycom/llm-app

Ready-to-deploy LLM application templates that provide retrieval-augmented generation (RAG) over live, changing data sources. The templates include built-in data indexing and search capabilities, removing the need to maintain separate vector database or cache infrastructure.

Project overview

Handles continuous, real-time synchronization of new data additions, deletions, and updates from connected sources, keeping RAG applications up to date without separate synchronization infrastructure.

Project type: RAG · Workflow · Vector Database
Use cases: Knowledge Q&A · Data Analysis
Deployment: Refer to project documentation
License: MIT

Best for

Developers, enterprise teams, and AI engineers building production-ready enterprise RAG and AI search applications that require continuous synchronization with live data sources.
Teams seeking to handle vector, hybrid, and full-text search directly in-memory without integrating separate external vector database modules like Pinecone or Weaviate.

Key capabilities

Templates automatically synchronize all new data additions, deletions, and updates from connected live data sources, such as Sharepoint, Google Drive, S3, and Kafka.
Provides in-memory built-in data indexing enabling vector search, hybrid search, and full-text search with cache, removing the need for external vector databases.
Uses GPT-4o during parsing to index PDFs and other documents, for extracting information including charts and tables from unstructured folders.
Provides a private, local RAG pipeline using Mistral and Ollama without external API dependencies.

Limitations and risks

Using templates with OpenAI GPT models incurs API costs and requires external network access.
GPT-4o and GPT models are optional but required for specific templates, such as the multimodal RAG pipeline. Mistral and Ollama models are used for the private RAG template.

Getting started

Setup is straightforward. Pick a template, run it out of the box using Docker containers or Python scripts, and connect to it via the HTTP API or Streamlit UI.

Alternatives and comparisons

Provides a unified Python API to store media, run models, index embeddings, serve endpoints, and version everything in a single system instead of maintaining separate glue scripts.
Provides an integrated framework to process and query multimodal document content across modalities for traditional text-focused RAG systems.
A lightweight, in-process vector database that runs locally without servers for production-grade similarity search with minimal setup.

Project comparisons

Evidence and sources

GitHub project description: Ready-to-run cloud templates for RAG, AI pipelines, and enterprise search with live data. 🐳Docker-friendly.⚡Always in sync with Sharepoint, Google Drive, S3, Kafka, PostgreSQL, re…
README: The Pathway Live Data Framework's **AI Pipelines** allow you to quickly put in production AI applications that offer **high-accuracy RAG and AI enterprise search at scale** using…
README: The apps can be run as **Docker containers**, and expose an **HTTP API** to connect the frontend. To allow quick testing and demos, some app templates also include an optional Str…
README: Basic end-to-end RAG app. A question-answering pipeline that uses the GPT model of choice to provide answers to queries to your documents (PDF, DOCX,...) on a live connected data…
README: [`Multimodal RAG pipeline with GPT4o`](templates/multimodal_rag/) | Multimodal RAG using GPT-4o in the parsing stage to index PDFs and other documents from a connected data source…

Reviewed at: 2026-07-17T06:41:43Z

Open original GitHub project