Skip to main content
POST
Create Data Sources
Add new data sources (YouTube videos, websites, Zendesk tickets/articles, etc.) to your Guru. The processing is asynchronous. Check the status of the data sources with the Get Data Sources endpoint.
PDF, Word (.docx), UDF (.udf), and Excel files should be uploaded directly through the Gurubase platform interface. We will add support for uploading these files via the API in the future.
If you do not have the relevant integration for Jira, Zendesk, and Confluence while adding a new data source, an error is returned.

Path Parameters

string
required
The slug of the Guru to add data sources to

Headers

string
required
Your API key for authentication. You can obtain your API key from the Gurubase dashboard.

Body Parameters

string[]
Array of YouTube video URLs to process
string[]
Array of website URLs to index
string[]
Array of Zendesk ticket/article URLs to index
For Gurubase Self-hosted version, Zendesk integration is currently in beta and available to selected users. For Gurubase Cloud version, Zendesk integration is available to all users.
string[]
Array of Jira Issue URLs to index
For Gurubase Self-hosted version, Jira integration is currently in beta and available to selected users. For Gurubase Cloud version, Jira integration is available to all users.
string[]
Array of Confluence Page URLs to index
For Gurubase Self-hosted version, Confluence integration is currently in beta and available to selected users. For Gurubase Cloud version, Confluence integration is available to all users.
object[]
Array of text sources to index. Each object should contain:
  • name (string): Name of the source
  • content (string): Text content to index
  • external_id (string, optional): Caller-owned identifier for the source. When you send the same external_id again, the existing source is updated in place instead of a duplicate being created. Reads (see Get Data Sources) return this id so you can reconcile against your own identifiers without keeping a mapping table. Must be 1-256 characters, start with a letter or number, and use only letters, numbers, dot, underscore, colon, or hyphen. If omitted, the legacy replace-by-name behavior applies.
  • context_prefix (string, optional): Text prepended to every chunk before it is indexed. Useful when a source is a transcript or fragment that needs a heading (for example a meeting date or a document title) so a chunk retrieved on its own still names its origin. Up to 4000 characters.
If you send a text data source without external_id that has the same name as an existing one, it will be updated with the new content.
Re-sending a source with an unchanged external_id, content, and context_prefix is reported as unchanged and does not consume the plan’s source quota, so a caller can safely replay its set on a schedule.

Response

The response is an array of processed data sources.
string
Type of data source (Website, YouTube, Jira, Zendesk, Confluence, etc.)
string
URL of the data source
This field is not present for data sources of Text type.
string
Processing status (e.g., “exists”, “success”)
number
Unique identifier for the data source
string
Title of the data source (can be null)
string
Content of the data source
This field is only present for data sources of Text type.