Data sources
Every source Dataki reads, and how it gets there
Six connectors in the app, and through them any hosted Postgres or MySQL. Google's products, Firestore and Stripe arrive through BigQuery, where they already export. Each route has a guide that sets it up, and every query runs read-only.
01A connector in the app
Connects directly
Sign in, or type the connection details once. A database, a warehouse or a sheet is queried where it lives, at the moment you ask, so there is no copy of it to keep current.
-
BigQuery A read-only service account in your project
-
PostgreSQL A read-only user, over TLS
-
MySQL A read-only user, over TLS
-
Supabase Sign in with Supabase, then add the password
-
Google Sheets Google sign-in, then pick a sheet
-
CSV and Excel Upload; each sheet becomes a table
Wherever your Postgres or MySQL lives
A hosted database is still Postgres or MySQL. Dataki connects over TLS whenever the
server offers it, which the hosts below require, and always from one address,
34.89.253.13, so a firewall can let in Dataki and no one else.
02Someone else's export, read in place
Through BigQuery
These have no connector in Dataki, and do not need one. Each already has a way into BigQuery, most of them free. Turn it on, connect the dataset, and Dataki queries it like any other warehouse.
- Your product
- BigQuery
- Dataki
-
Google Analytics 4 GA4's free daily export: every event, unsampled. Guide · about 10 min
-
Firebase Firebase's export: Analytics, Crashlytics, Performance, Messaging. Guide · about 5 min
-
Google Search Console The bulk data export: every query and page, no row cap. Guide · about 10 min
-
Cloud Firestore Google's extension streams each change as it happens. Guide · about 15 min
-
Stripe BigQuery's own Stripe connector, free while in Preview. Guide · about 15 min
03When there is no live route
Through a file, or a sync
An export you upload
A CSV or Excel export is queried with full SQL, each sheet as a table. It is a snapshot, so it suits a question now rather than a dashboard that should keep itself current. MongoDB works this way today, or through a copy in BigQuery.
MongoDB guideAnything a sync can land
If a tool can put it in BigQuery or Postgres, Dataki can read it there. BigQuery's own Data Transfer Service copies Salesforce, HubSpot, Shopify, Google Ads and more, some in Preview. Airbyte, Fivetran and dlt land hundreds of others in either.
04No connector of their own
Not yet
The connectors people ask for most. None has a date, so we will not pretend otherwise; until one ships, the sync route above is the way in.
- HubSpot
- Salesforce
- Shopify
- Zendesk
- Intercom
- RevenueCat
Need one sooner, or something only your company uses? We build custom connectors for partners, typically in two to four weeks.
Free while we are in beta
Connect a source. Ask it something.
One source is free. It takes about five minutes and a read-only connection, and the answers you keep become dashboards with links that stay live.