Github

🤖DataFlint Copilot

This feature is available in DataFlint SaaS Only

Summary

DataFlint Copilot is a vscode/cursor extensions that make your IDE to be production-aware. You can see code highlights of your last production run, and optimized the spark code based on DataFlint actionable insights.

It's powered by the MCP protocol, which enable LLMs to get additional context. In DataFlint SaaS case it means access to all your applications, all your spark job runs and an enriched SQL plans of all of your queries.

Installation takes just few clicks, and should take just few minutes!

Installation guide

  1. In Cursor, go to Extension tab:

  1. Search for DataFlint and press "Install"

  1. DataFlint icon was added to the status bar, press on it (or alternatively, run the command "DataFlint: Login")

  1. You will be get this popup, press Open (and also suggested to set "Configure Trusted Domains" and press enter afterwards)

  1. In the link you will log in using your Google OAuth, afterward you should get this screen:

  1. Return to Cursor, you should see a DataFlint "connect" icon, meaning you authenticated successfully

  1. You will get a popup to enable dataflint-copilot MCP server, press "enable". It can also enabled by going to settings -> tools and integration -> MCP Tools -> dataflint-copilot -> enable toggle

After this step, your AI IDE is connected to your DataFlint SaaS account via MCP and you are ready to go!

Copilot Usage

Now, when you are using the Cursor chat window and add a spark job as context, press "/" and you will get 2 dataflint-copilot prompts "highlight" and "optimize":

Highlight will add code highlight to your current job based on the last run. You can also trigger it by asking Cursor "add highlight for this job"

Optimize also optimize the code based on DataFlint alerts. you can also trigger it by asking Cursor "optimize this job".

Last updated