Getting started

You need an organization, a team, and a project before anything else will stick. Datasets do not live at company level.

First hour

1. Accept the invite and pick the team in the header. The wrong team is the usual reason “my project vanished.”
2. Open (or create) a project named for the campaign (cable-qual), not for a single file.
3. Create a dataset, upload a Touchstone or CSV, and wait until traces such as S21 appear.
4. Set parameters in the same sitting (lot, temp, sn, instrument). The filename is not a catalog.
5. Open the auto-created notebook and confirm the overlay is the measurement you meant.

If you work in Python as well, install the invited-beta wheel (not PyPI), run cloudsprite init (or set CLOUDSPRITE_API_KEY), then cs.connect(), cs.set_org, cs.set_team. Full walkthrough: Quickstart; the loop itself: Python SDK how-to.

To ask the assistant from Claude, add the public plugin (cloudsprite-io/cloudsprite-plugin) and sign in. See Assistant and Claude plugin.

Mental model

organization → team → project → dataset → traces
                              → notebook
                              → scripts (optional)
                              → workflows (optional)

A dataset is one measurement or one published computation. A trace is S21 (or similar) on that dataset. Notebooks plot traces, not “whole datasets.”

Roles

Viewers can inspect. Members can upload, tag, publish, and run scripts they can reach. Admins/owners manage membership and settings. Details: Organizations, teams, and roles.

What not to do on day one

• Do not invent REST paths. Use the app or cs.projects / ds.upload / cs.publish.
• Do not paste API keys into chat.
• Do not skip parameters because “it’s in the filename.”