← Browse docs

Start here

Quickstart

Sign up, import a dataset, and build your first cross-tabulation. Five minutes, end to end. Everything beyond this is optional until you need it.

Import a dataset #

Drag an SPSS .sav or CSV file onto the drop zone.
Your file is parsed right in your browser — survey data never touches Recense servers.
SPSS files load directly. CSVs show a review step where you confirm column types.
Recense then offers optional Smart prep: the agent reviews labels, types, and question groups. You can attach a project brief or the survey questionnaire (PDF) first so it has more context — or skip it.

Explore the dataset #

After import, the Dataset view shows your variables, question groups, and raw data.
Check variable types, labels, and groupings. Fix anything that looks wrong here — it feeds everything downstream.

Build your first table #

Switch to the Canvas view.
Drag a variable to Rows and another to Columns in a new table.
Your cross-tabulation appears instantly — all computation happens right in your browser.
Use the toolbar pills to add measures (percentages, counts), apply weights, or enable significance testing.

Set up AI (optional) #

Pick one of three ways to work with the agent:

Your own key (BYOK): add an Anthropic, OpenAI, Google, or Fireworks key in Settings → Provider keys.
Built-in: activate the hosted agent included with your plan in Settings → Hosted agent.
MCP: connect Claude Code, Claude Desktop, or ChatGPT desktop using an API key from Settings → API access.
The chip at the top of the agent panel shows which route the agent is using — e.g. "Your key · Claude Sonnet 4.6" or "Hosted" — and switches between them.

Save your work #

Recense autosaves to the browser every 30 seconds.
For a portable copy, use Save & export → Save to file to download a .recense file.
Cloud save is available on Pro plans.

Next steps