What is redis-ai-research-public?

Redis AI Research is a public collection of open-source projects from the Redis AI Research team. The repo explores how Redis can be used as a backbone for modern AI applications. The projects inside focus on practical patterns like semantic caching, vector search, retrieval-augmented generation (RAG), and multi-agent orchestration. It's essentially a showcase of reference implementations and experiments demonstrating what you can build when you use Redis alongside large language models. Each top-level folder is a standalone project. For example, one folder contains a coding-agent service that can spin up isolated workspaces for different users and run tasks in parallel. Another project lets you ask questions about a dataset in plain English and generates data analysis code to answer them, using Redis to remember past mistakes and avoid repeating them. There is also a benchmarking project that tests how fast Redis can search through hundreds of thousands of chat histories using different storage setups. The opencode-spec-optimization project automatically rewrites agent instructions until the agent reliably passes a set of measurable goals. Each project has its own README with specific setup steps. This repository is built for developers and technical founders who want concrete examples of how to wire Redis into an AI stack. If you are building an AI agent and need it to remember context, cache expensive LLM responses, or search through documents by meaning rather than keywords, these projects serve as working blueprints. The examples are practical, you might use the data-analysis agent as a starting point for an internal business intelligence tool, or adapt the coding-agent service to build your own automated code-review system. A notable aspect of how these projects are structured is that they are split into self-contained directories rather than one tightly coupled application. Most of them assume you have a recent version of Redis running and a Python environment set up, along with API keys for whichever LLM provider the specific project uses. The README does not go into deep detail on the inner workings of every project, so you are expected to dig into the individual project folders to understand the full architecture and tradeoffs of each implementation.