On this page
Connect an HTTP or HTTPS Data Source
A Data Source saves a connection to a supported external dataset and creates a refreshable MiniUp snapshot. Use it when an app needs a reusable copy of a web dataset instead of uploading a new file by hand each time.
Create an HTTP Data Source
- Open Dashboard → Data Sources → New source and select the HTTP/HTTPS source type.
- Enter Source name and choose the Destination MiniUp site.
- Enter the direct HTTP/HTTPS file URL, not a webpage containing a download button.
- Choose Format: auto, CSV, JSON, JSONL, NDJSON, GeoJSON, Parquet, or GeoParquet.
- Choose Authentication: None, Bearer token, or API key header. Enter credentials in the dedicated fields. Supported API-key header choices are shown in the form.
- Continue to Test connection and select Save and test.
- Use Preview source to inspect schema and sample records without creating an output dataset.
- Choose Create snapshot after selecting the destination Site.
Refresh and inspect a Data Source
Open the source detail page to test, preview, or refresh it. Inspect Refresh and connection history, the latest result, row/byte counts, warnings, and inferred schema. Update connection settings when the source URL or credentials change. Use the available disable/delete controls when a source is no longer needed; review the confirmation before deleting. Deleting a Data Source leaves its latest output dataset available; remove or restrict that dataset separately if it should no longer be used.
Understand snapshot behavior
A snapshot represents the data from a completed import, not a continuous live connection. Refresh it when you need updated data. A failed refresh leaves the last successful snapshot available. Check the latest successful run before claiming an app shows current upstream data. Do not assume automatic scheduled refresh unless the product explicitly provides that setting.
Example and common problems
Use https://example.com/operations.csv as a direct source URL for an operations dashboard. If preview fails, verify the URL returns a supported file, credentials are valid, and the file fits the current import limits. URLs containing credential-bearing query parameters are not supported; use the dedicated authentication fields.