How it works
- In Data sources, choose Add source, then REST API.
- Give the address of the list you want, for example
https://api.example.com/v2/orders, and how to sign in. - Klyrova reads the first page and shows it to you as a table before anything is saved. If it looks right, connect.
- When you build a dashboard on it, tick its table and choose how the dashboard reads it, as for a database: live, or imported (on Pro and Teams) at set times, up to eight a day.
Signing in
No sign-in for open APIs; a key in a header (such as X-API-Key); a key in the address; a bearer token; or a user name and password (basic authentication). Keys are encrypted, tied to your account, and never shown again.
Finding the records, and every page of them
Klyrova finds the records by itself: the answer, when it is a list, or else the biggest list of objects one or two levels in (the data or results of most APIs). You can also name the path. It follows the API's paging to the end, in whichever way the API pages:
- a next address or a cursor in the answer;
- a
Linkheader; - a page number, or an offset and a limit.
Rate limits are respected: an answer of "too many requests" or "unavailable" is waited out, then the reading goes on.
From JSON to a table
Each record becomes a row. Nested objects are flattened into columns such as customer_city, three levels deep; a list of plain values goes in one cell; anything deeper is kept as JSON text. Times with a time zone are stored in UTC. Columns get types (numbers, dates, true or false, text) from their values, so charts can add up and sort them.
Ask and build
Once connected, Klyrova's agent can query the table in the chat, and you can build dashboards and reports on it, or join it with your database in a data model. Read live, every refresh, filter and question asks the API again and Klyrova keeps nothing; imported, the dashboard's data model keeps a copy that each import replaces.
Safe by default
Klyrova only sends GET requests to public HTTPS addresses. It refuses private, loopback and link-local addresses, checks every redirect the same way, and never forwards your key to another host. Each answer is capped at 64 MB; a live read at 250,000 rows a table, 400 requests and a minute; an import at two million rows and 400 MB a table, 3,000 requests and 20 minutes.
Questions about REST API dashboards
Does the API have to return JSON?
Yes. Klyrova reads JSON answers; an HTML or XML page is refused with a message saying so.
Can Klyrova call APIs inside my company network?
No. It only reaches addresses on the public internet over HTTPS.
Can it write to the API?
No. It only reads, with GET requests.
Does it handle pagination?
Yes: next links, cursors, Link headers, page numbers and offsets, until the last page.
How often is the data refreshed?
Read live, each time a dashboard, a filter or a question asks. Imported, at set times up to eight a day on Pro and Teams; Import now refreshes at once.
Updated 16 November 2026
