Failing to abide by the MIT license is copyright infringement. My advice is to contact these guys: https://softwarefreedom.org/ They likely can file a cease and desist on your behalf.
However, I took a closer look at the files in question. The MIT license requires that they retain and provide copyright notices, but you never put copyright notices in your files. The only place where you appear to have placed a copyright notice is in the LICENSE file:
https://github.com/spegel-org/spegel/commit/23ed0d60f66dd292...
Things become interesting when I look at their LICENSE file. They appear to have tried to relicense this to Apache 2.0 before backpedaling and reinstating the MIT license:
https://github.com/Azure/peerd/commit/473a26c808907f2d9f7b7f...
Unless they forked from a very early version of the project that did not even have the LICENSE file, they removed the sole copyright notice you had in the repository. That brings us back to my original thoughts, which is that they have committed copyright infringement, and you should contact OSS friendly lawyers about it.
I am not a lawyer, but I do contribute to various OSS projects and all of the ones to which I have ever contributed have copyright notice headers at the top of every file to ensure proper attribution is maintained no matter where that code is used. Beyond having that sole missing copyright notice reinstated, I am not sure what else you could expect since none of your files have proper copyright headers in them. The SFLC guys would be in a better position to advise you, as they are actual lawyers.
File a police report, go through the right channels. If you know its yours, call the police department non-emergency and explain the situation
There is hypocrisy here around internet archive, it's totally OK to store copy-write content on the archive, but its not OK when a company does so on their own.
I think it's a bit discourteous to shoo Daniel away due to an out of scope
report, then cry wolf when your clients do actually feel that this warrants
a response. The fact that you made changes to your systems in response
indicates that this wasn't as benign as it first seemed.
IMO Zendesk should do the right thing and issue a reward. An issue was
reported and ultimately resolved in some fashion. Continue to encourage
researchers to bother reporting things to you. Yes, you have a little egg
on your face due to the end-run via your clients, but that's life, Zendesk
will survive.
Its transcribing the meeting but hallucinates badly... both in fast and thinking mode. Fast mode only transcribed about a fifth of the meeting before saying its done. Thinking mode completely changed the topic and made up ENTIRE conversations. Gemini 2.5 actually transcribed it decently, just occasional missteps when people talked over each other.
I'm concerned.