Your first report

You need a terminal and a browser. A coding agent helps write the report, but you can try the whole rendering path yourself first.

Get the command

The native release is not available yet. Try the playground now. The walkthrough below is for an installed native command.

Make the smallest useful report

In a fresh working folder, run:

richview init first-report
richview validate first-report/report.rmd --format json
richview check first-report/report.rmd --format json
richview generate first-report/report.rmd --format json

Open first-report/Rendered/report.html in your browser. It is a complete interactive HTML report in Sleek, titled Direct visits lead our traffic, with a 910-visit metric and fictional observations in data/visits.csv. Hover or focus the chart, inspect its table and try a download. Keep the source beside the result: you edit report.rmd, then generate again.

validate checks syntax, schema and local resources. check also renders the report and inspects chart geometry. In JSON output, status: clean means no errors or warnings; ok: true can still include warnings. Read those before treating the report as finished.

Keep a preview open

richview serve first-report/report.rmd --port 4711

Open http://127.0.0.1:4711/. Save a source or data edit and the preview reloads. Stop the command with Ctrl+C. The server binds to loopback by default; this is a preview of one local report.

Give your agent the job

Paste this into Claude Code, Codex, Cursor or another coding agent that can read your files and run the command:

If RichView is installed, run richview learn. Then make me a beautiful report from my data, or from the starter example if I have none. Check it, fix anything it flags, and show me the result. Keep everything on my machine.

Attach or name your data files and explain the question you want the report to answer. The agent writes ordinary saved Rich Markdown and data files. You can read them, put them in Git and keep working after an interrupted session. Review the interpretation as well as the chart.

Take a finished example apart

The orders example uses a different fictional dataset to demonstrate a metric, table, caveat and local styling. Download its source, style YAML and CSV into this layout:

orders/
  report.rmd
  report.rstyle.yaml
  data/orders.csv

Keep exactly one .rmd in each report folder, including its subfolders. Store source backups outside that folder or use Git. Generated files go into Rendered/ with a capital R.

Next: use the CLI, choose a style, or learn the source format.