Skip to main content

Weaviate development setup

This page describes how to run Weaviate from source (git checkout / tarball) locally.

tip

You can find the source code at the Weaviate repo.

Prerequisites:

Running from source​

The fastest way to run Weaviate from source is to issue the command below:

tools/dev/run_dev_server.sh <configuration>

Where <configuration> is one of the server configuration ($CONFIG) values in /tools/dev/run_dev_server.sh. For example, you can run:

tools/dev/run_dev_server.sh local-openai

To run the server locally with the OpenAI module.

The default configuration is local-development which will run the server locally with the text2vec-contextionary and backup-filesystem modules.

You can also create your own configuration. For instance, you can clone an entry (local-all-openai-cohere-palm is a good start) and add the required environment variables.

Running with Docker​

To run with Docker, start up the Weaviate container and the container(s) for any additional services with

tools/dev/restart_dev_environment.sh [additional_services]

then run the development server as described in the section above.

For example, the setup below uses Docker Compose to spin up Prometheus and Grafana instances. Those are pre-configured to scrape metrics from Weaviate. Using this setup, you can:

  • access Weaviate on port 8080
  • access Grafana on port 3000 (Login: weaviate/weaviate)
  • if necessary for debugging - access prometheus directly on port 9090
tools/dev/restart_dev_environment.sh --prometheus && tools/dev/run_dev_server.sh local-no-modules
info

This setup is for contributors to the Weaviate code base. If you are an end-user of Weaviate looking for a Prometheus-enabled example, please see this documentation page or this example.

Below are more examples of running Weaviate with Docker.

Transformers t2v only​

tools/dev/restart_dev_environment.sh --transformers && ./tools/dev/run_dev_server.sh local-transformers

Contextionary t2v & Transformers QnA​

tools/dev/restart_dev_environment.sh --qna && ./tools/dev/run_dev_server.sh local-qna

The above commands are subject to change as we add more modules and require specific combinations for local testing. You can always inspect restart_dev_environment.sh and run_dev_server.sh to see which options are available. The first option without any arguments is always guaranteed to work.

To make queries from a web interface, use the WCS console to connect to localhost:8080.

More Resources​

For additional information, try these sources.

Questions and feedback​

If you have any questions or feedback, let us know in our user forum.