Skip to content
Documentation

Sources and connectors

Everything in the Library starts as raw data from somewhere: a spreadsheet on your laptop, a database, the systems your business already runs on. Sources and connectors are how that data gets in, and how it stays current.

Add data to Querri 1:30 Upload files, connect a database or a business app, and find it in the Library.
Read the transcript

Why bring your data into one place

Sales in a spreadsheet, customers in your CRM, orders in another app. Scattered data slows down every answer. Querri brings it all into one Library, so your chats, projects and dashboards work from the same data. If you look after your team’s data, this is where you start: click Connect in the left menu.

Upload files from your computer

That opens Add data. To upload from your computer, click Browse. Querri reads CSV, Excel, JSON and Parquet files. Or simply drag your files straight onto the page.

Connect a data source

Or connect a data source. Search by name, or filter by category, from databases to CRM and accounting apps. Anything already set up shows Connected.

Set up a database connection

Pick a source to open its setup. A database like PostgreSQL asks for its host, username and password. Then test the connection, and Querri lists the databases it can reach. Apps like HubSpot open a sign-in window instead.

Find your data in the Library

However you add it, your data lands in the Library. Open Files to see your files and tables. When the status says Ready, it’s done processing. Now ask Querri a question in plain English, and it answers from your data.

The idea

Section titled “The idea”

A source is one table or file you’ve connected or uploaded. Views are built from sources, KPIs are measured from them, and facts get attached to them.

A connector is the link to a system Querri pulls data from. One connector can bring in many tables, and each table it loads becomes a source.

Add data puts the two ways in side by side. You can upload files from your computer, or drop them anywhere on the page. CSV, Excel, JSON and Parquet files become data you can ask about, while PDFs, images and documents join the Library as files. Or you can connect a system, like a database, a CRM or an accounting package. You don’t need to be an admin to start a connection there, and admins also manage connectors in Settings under Connectors.

Pick the workspace before you add anything, because that’s where the data lands. A new connector goes into the workspace you’re in if you’re a member, and into your private workspace otherwise. Files you upload in the Library go to the workspace you’re viewing. If you can’t add files there, they’re saved to your private workspace and a message tells you. Workspaces and privacy covers the rest.

Querri works from a copy

Section titled “Querri works from a copy”

When a connector syncs, Querri copies the data into its own storage and answers from that copy. Answers are only as fresh as the last sync.

So how does it stay fresh? A connector’s owner can give it a schedule. Full Refresh reloads everything, and Incremental fetches only new or changed records, tracked by a cursor column such as updated_at. The presets run from every hour to every month, and a schedule that fails three times in a row is switched off until someone fixes it. You can also refresh by hand. After any sync, the views built on that data rebuild by themselves.

Scheduled syncs use credits. Syncs you start yourself, and file uploads, don’t.

When data arrives

Section titled “When data arrives”

Querri measures each new source as it comes in: its columns, their types and some example values. When two tables share a key, it records the link between them on its own.

Once newly connected or uploaded data settles, Querri links it to the views you already have, marks the questions it can now answer, and may leave View to build cards in your Inbox. It doesn’t build views by itself. The nightly pass takes a second look later, measuring column health and spotting the same customer or product living in two systems, as How the Library learns explains.

One warning: disconnecting a connector deletes the data it synced, along with the access people had to it.

An example

Section titled “An example”

Curio, the sample company, comes with data from the kinds of systems a business like it runs on: finance and accounting, a CRM, an online store, point-of-sale in its shops, ad platforms and shipping. They arrive with the demo data. They aren’t connections to set up, and some of them aren’t in the list of connectors you can add yourself.

The sample is realistic rather than tidy. Its own description says “the same customer appears in several systems under different keys”. That’s the mess sources carry in, and it’s why the view Golden customer record v2 exists: one row per customer across four systems, matched on email first and then phone.

Open a source and you can check it came in right. Curio’s in-store transactions table has no customer ID on a large share of its rows, so the Columns list shows that column’s filled figure in red. You know before you ask that a per-customer breakdown of store sales will be incomplete.

Where it lives in Querri

Section titled “Where it lives in Querri”

Connect in the app rail opens Add data, with Upload from your computer at the top and Connect a data source below it. The Library’s + button offers Upload files… and Connect a data source too.

In the Library, the browse panel’s Connectors section lists each connector. Right-click one for View assets, Re-sync now, Configure… and Remove connector. Sources appear on the Files tab with the type Dataset, and opening one shows when it last synced, where it came from, its columns and what uses it. Table and column panels walks through that panel, and the Connectors overview lists the connectors you can add.

Next steps

Section titled “Next steps”