datajoi. Docs
Docs

Getting started

From a first source to your first answer with a receipt.

This walk-through uses the sample Northwind workspace on the live demo. Everything works the same on your own workspace.

#1. Sign in

Use your work email. You receive a one-time link, or you can continue with Google when your organisation has enabled it. Workspaces have three roles: owner, editor and viewer. Viewers can read and ask; editors can propose and run SQL; owners can change settings and trust.

#2. Connect a source

Open Connect and choose Add source. Files, databases (Postgres, Google Cloud SQL), REST APIs, Cloudflare D1, KV and R2, Google Cloud Storage, Amazon S3 and Google Sheets are supported. Credentials are stored as named variables, never in the page. See Connect.

#3. Approve the schema contract

An agent introspects the source and files a schema contract in your Inbox: column types, PII flags and a sample. Review it, edit it if you need to, then approve. Only after approval does the pipeline load data into your lake.

#4. Let the agents work

Quality suites, transformation models, documentation and semantic terms each arrive as further proposals. Approve the ones you want. The dashboard batches what needs you, for example "Draft 5 test suites".

#5. Ask a question and read the receipt

Open Data Studio. Ask in plain English, or write SQL yourself in the editor. Each answer shows the SQL, the tables it read, their freshness and the semantic-layer version it used.

#Next