From a new project to a published dashboard.

Sign up & create your project

Sign in at overscore.dev with Google. Click Create project and pick a name — that becomes your subdomain.

https://your-project.overscore.dev

Choose your data

During project setup, choose Choose a dataset to add a real starter dataset and follow the in-app walkthrough. You can also download retail sales, bike rentals, or penguin measurements and upload the file in your project’s Datasets page. Sources and credits.

Using your own data instead? Upload a CSV, JSON, or Parquet file, or open Connections in your project to connect BigQuery, Snowflake, or Postgres.

Create a dashboard and open it locally

Create a dashboard in the Hub, then copy the command from its Overview. It creates or updates the local project folder and loads the Overscore context your agent needs.

$ npx @overscore/cli open sales --project your-project

Build with your AI agent

Open the project with Claude Code, Codex, Cursor, or the coding agent you already use. Describe the dashboard you want. The local project includes the Overscore rules, structure, and data context it needs. You still review the generated code and SQL before it runs against your data.

Deploy with one command

When the dashboard looks right, run the Overscore deploy command. Built files and source go to our infrastructure as a versioned deployment. Open the dashboard from the Hub to review the active build before you publish it to an audience. Full CLI reference →

$npx @overscore/cli deploy --message "first dashboard"

Share with your team

Publish the dashboard, choose its audience, and send the URL. Private viewers sign in with Google; public dashboards can be opened without an account. Result freshness follows the project's cache and refresh settings. Eligible paid projects can also offer the dashboard as an installable app. Auth & permissions docs →