eCommerce Demo    Docsearch Demo    Redis VSS Documentation   

Getting Started: Redis Vector Similarity

This entry-level tutorial is meant to guide you through:

• Setting up a RediSearch + Jupyter + Streamlit docker stack
• Creating text-based embeddings (vectors) from a sample/toy dataset
• Storing vector data in RediSearch
• Running vector queries and hybrid queries
• Hosting a simple Streamlit UI

---

WARNING -- this is a toy example/demo. This is not meant to replicate production in any way. Rather, use this to learn the basics to apply to your own data and pipelines.

Docker Setup

Make sure you have Docker Desktop installed on your workstation.

Run this command to start up the stack of services including Redis and Jupyter:

$ docker compose up

Run this command to tear down the stack:

$ docker compose down

If at any point you need to trouble shoot, run this command to check running docker processes:

$ docker ps -a

Check logs of a docker container with the id found in last step:

$ docker logs {CONTAINER_ID} -f

Jupyter Notebook

We use Jupyter notebooks here to guide through the tutorial of loading data, creating embeddings, storing in Redis, and creating a search Schema.

After running the docker compose up step, you should see a link in the logs like this:

Use the last URL listed... the one that has your own custom token -- not this exact one :)

Streamlit

We've included a sample Streamlit UI that allows you to enter a search query and explore a subset of documents with AI-powered vector similarity search.

The UI can be extended or modified to fit your schema and usecase. This just gives you a starting point!

WARNING: Streamlit app only works once the data has been loaded to Redis. Use the Jupyter notebook for that (see above). Refresh the app once dataset is in Redis.

Full Example

See our arXiv paper search demo/example that includes a full front end and backend system if you're ready for the next step.