To get started, head over to https://dashboard.onvo.ai to create an account. Once an account has been created you can go to https://dashboard.onvo.ai/integrations. Here, click on New Integration on the top right corner.
Now you can choose from our list of basic integrations to connect to.
Click on connect next to the card of the integration you wish to connect.
File based integrations
For file based integrations like Excel and CSV, fill in a name for the integration and upload the file. Once the file has been uploaded, click on Connect.
Make sure the file completed uploading before clicking Connect or else your integration will not work.
OAuth based integrations
For OAuth based integrations like Airtable and Google Sheets, fill in a name for the integration and click on connect. This will take you to the respective platform to authenticate.
For Google sheets, you may see a warning saying this app is not trusted, you may have to click on Advanced and to trust this app.
Databases and data warehouse integrations
For databases and data warehouse integrations, you just have to fill in the credentials and click on connect. We recommend creating a separate user account with read access to just the tables you want to expose (for more information about security, head over to https://www.onvo.ai/payload/docs/securing-your-data-on-embedded-dashboards). If your database supports IP whitelisting, you have to add 49.13.36.246 to be able to connect to your database.
Your database / data warehouse has to be accessible on the public internet. If it is only available on an internal network, we will not be able to connect or read data.
Auto discovery of datasources
Once you click connect, Onvo AI automatically attempts to discover all the datasources in your integration.
- For csv files, this would be a single datasource
- For excel, this would be each sheet in the file
- For Airtable and google sheets, this is all the documents and each sheet in each document
- For databases and data warehouses, this is all the schemas and each table in each schema
The auto discovery process will find all the tables, count the rows and columns in each one and automatically annotate a description for each one to build up a context of how your data is structured and linked. If there are tables that have errors (For example: an excel file that has merged cells, multiple tables in a single sheet or elements like charts), this will be marked as errored out and you will not be able to create datasources out of them.
Now your integration is ready to use and you can start create datasources from them
Last updated on Sep 12, 2025