A Viewer URL names a location in the Viewer, also called a route: a recording, a page, or something on a Rerun server. Opening a Viewer URL goes to that location.
| URL | Location |
|---|---|
about:recording?app_id=<APP_ID>&recording_id=<RECORDING_ID> | A recording that is open in the Viewer |
about:table?id=<TABLE_ID> | A table that only exists in the Viewer |
about:chunk_store?app_id=<APP_ID>&recording_id=<RECORDING_ID> | The chunk store browser of an open recording |
about:settings | The settings page |
rerun+http://<HOST>:<PORT> | The catalog of a Rerun server |
rerun+http://<HOST>:<PORT>/entry/<ENTRY_ID> | A table or dataset on a Rerun server |
rerun+http://<HOST>:<PORT>/dataset/<DATASET_ID> | A dataset on a Rerun server |
rerun+http://<HOST>:<PORT>/folder/<PATH> | A folder of entries on a Rerun server |
Query values are URL-encoded, so a recording with the application id my app has a URL like about:recording?app_id=my+app&recording_id=….
Use rerun+https:// for a server with TLS.
A dataset URL with a segment_id names a recording on a server.
See Generate segment URLs in dataframes for those, and for the #when=… fragment that sets the time cursor.
In the web Viewer
The web Viewer shows its Viewer URL in the address bar, URL-encoded as the url query parameter of the page:
https://rerun.io/viewer?url=rerun%2Bhttp%3A%2F%2Flocalhost%3A51234%2Fentry%2F<ENTRY_ID>Opening such a page opens the URL in its url parameter, and a page with several url parameters opens all of them.
This makes the address bar a link to share, except for about:recording, about:chunk_store, and about:table: they name a recording or table by its id, and only work in a Viewer that already has it open.