At this moment, there's one called Status (which seems to be new) and one for uptime.invidious.io (which has been there for a while, had to blacklist it in my userscript)

As per title, the instances API responds with error 500 or with empty results

https://github.com/iv-org/instances.invidio.us/blob/master/src/instances.cr#L169

suggestion (fix?): make it so that e.g. ?sort_by=health,users-reverse lists them first by health and then by least users (right now it lists them first by health and then by users and then reverses the entire list)

I would love to be able to filter the json file:

Actually, this https://instances.invidio.us/instances.json?sort_by=type does not work.

which is the URL in the brackets after yewtu.be on the instance list page

https://yggdrasil-network.github.io/about.html

Is there a way to know whether an instance supports/accepts API calls before making the actual call? As an example, I used to use invidious.snopyta.org, but then it started blocking API calls, replying with 403.
Maybe it's me who's doing something wrong, I'd also like to know if that's the case.

And to make it clear, I'm looking for a programatic way, maybe through api.invidious.io/instances.json?

Hello,

When retrieving the instances JSON with ?sort_by=health, the server returns 500 with Cast from Float64 to String failed, at /usr/share/crystal/src/json/any.cr:250:5:250. Same error occurs after clicking on "health" on the instances table.

This started happening just today (22.04.)

invidious.ggc-project.de - https - - 🇩🇪 DE 99.516
yt.maisputain.ovh

my instance (invidious.namazso.eu) does not allow registration, yet https://api.invidious.io/ lists it as one that does

Is there any plan to revive the monitor data previously provided by the api? Currently instances.json is returning null on monitor:

... "type": "https",
"uri": "https:// yewtu. be",
"monitor": null ...

where as previously it provided information like "success":

... monitor: {
monitorId: 782758961,
createdAt: 1558969602,
statusClass: "success", ...

I have removed the criteria from my response parsing, but would like to know if I should make it permanent

Note: Originally opened in main issues list and moved per request.

Hello! There's an issue I'm having with CORS while developing my project

Now, I'm not an expert on CORS, but I think this is the issue:

Most browsers (Firefox and Chromium-based ones afaik) does not only check for Access-Control-Allow-Origin but a few more headers while doing a CORS preflight request:

• Access-Control-Allow-Methods
• Access-Control-Allow-Headers (probably optional)

I propose that instances-api checks for the Access-Control-Allow-Methods header while checking if an instance has CORS enabled.

Should be a one liner:
https://github.com/iv-org/instances-api/blob/master/src/instances.cr#L79

Sadly invidio.us is currently down. So it would be great, if there would be an automated forwarding to another more healthy instance, of which the server is the closest to the requesting IP location

The Invidious API is currently not returning the list of instances; instead, it is returning an empty list. Additionally, the page in the documentation that lists all the instances is returning a 404 error.

Can you add a filter argument to the API that I can filter out instances that don't have CORS set to *?
Or a filter for instances that allow the usage of the API.

Currently onion and https instances are mixed
https instances should be at the top of the instance list

Is your enhancement request related to a problem? Please describe.

https://redirect.invidious.io/ lists public instances of the site sorted by health but even sites with 100% health can sometimes be very slow compared to other instances. Even if you select the top instance you can end up with a slow one.

Describe the solution you'd like

I wish it would sort by health and speed of the site. There should be another column indicating how fast the site runs and how well it plays and loads videos. Some instances take longer than others to load videos. Health only seems to refer to how little the site crashes (?).

Describe alternatives you've considered

Let the user have to test all kinds of instances before they might find one that they find fast enough. Personally I remember yewtu.be to have an acceptable speed so I usually choose that but there might be better ones.

Describe the bug
incorrect spelling of url in https://api.invidious.io/instances.json

Steps to Reproduce

1. Go to https://api.invidious.io/instances.json?pretty=1&sort_by=type,users#:~:text=%22uri%22%3A%20%22https%3A//yewtu.be%22%2C
2. See error

Screenshots

It seems Access-Control-Allow-Origin directive is missing on the server and disallows websites like https://booteille.gitlab.io/invimatrix/ to work.

As an example, the same code was used for https://booteille.gitlab.io/peertubatrix/

Trying to open YT video opens a list of proxies where one is yt.drgnz.club displayed as 100% health, yet it is false because it is not working, because that site https://yt.drgnz.club/watch?v=abcdef says:

Can't load the video on this Invidious instance. YouTube is currently trying to block Invidious instances. Click here for more info about the issue.

Please do not show instances which has this "issue" and can not load the video (currently). User is only mislead by the 100% health wasting the time.

Invidious CI recently moved from Travis to Github Actions for CI.

This project would also benefit from CI (using the Github Actions too).

The file should be pretty similar to the one used in Invidious since it's also Crystal.

@saltycrys since you're the one that did the Invidious CI, you might be interested in doing that.

I am referring to the table found here:
https://api.invidious.io/
But it would be nice to also have this feature in this list: https://docs.invidious.io/instances/

What I mean is that full country names should be accessible to people who are searching for instances, as it is more convenient than manually looking up each country code (or being good at recognizing flags). This could be done by either adding the full name as a row in the table, or as a hover tooltip when hovering over a cell in the location column.

I sometimes get an empty page from instances.invidio.us.
I guess it's because sometimes, uptime.invidio.us time out.

To avoid to break everything when such cases happen, I suggest to cache good results from uptime.invidio.us for some times (and invalidate the cache after some minutes if we get good results again).

That would also reduce the number of calls made on the uptime server since it should become more and more used by third parties.

Hi, I wrote a fork of Privacy Redirect that uses your API to retrieve only working instances of Invidious.

Lately I've noticed redirects no longer work, and the cause seems to be that the UptimeRobot status page that the API requests is broken (here and here).

Do you know what's causing this, and will it get fixed?

Describe the bug
incorrect spelling of url in https://api.invidious.io/instances.json

Steps to Reproduce

1. Go to https://api.invidious.io/instances.json?pretty=1&sort_by=type,users#:~:text=%22uri%22%3A%20%22https%3A//yewtu.be%22%2C
2. See error

Screenshots

What does 99.553% health mean?

Due to the addition of multiples Invidious instance violating the AGPL. The documentation has been moved to its own repository: https://github.com/iv-org/documentation

This project needs to be updated to use the new document: https://github.com/iv-org/documentation/blob/master/Invidious-Instances.md