I've bought a few well known transcription deals here, and tested Laxis against them with the same uploaded files.
In every case, Laxis' accuracy was noticeably better. So they definitely have a great platform to build upon, and they have what matters most, nailed.

The recent Ai assistant feature is shockingly good and all three features (summary, client request and post-meeting email) work very well.
Again, more solid features to build upon.

The plan limits are exceptionally generous.

The only things letting Laxis down are a few quality of life features and a lack of speaker identification, which I consider important.

Speaker identification is very important, but I can understand why they removed it, as poor speaker identification can cause more trouble than it's worth. Regardless, I think it's important enough to keep and try improving, and some of those issues can be mitigated with UI improvements.

The UI itself is rather basic, the color scheme is dull, and screen space could be used far more efficiently.

With regards to speaker identification, improvements in transcript editing would help immensely. Currently, I have to name every single speaker, over and over again.
With many other transcription apps, I can rename "speaker 1" to "John Smith" and this will ripple through the entire transcript.

Also, the "meeting templates" function is, while a great idea, difficult to utilise fully, as I have to search through transcripts for telltale phrases that trigger recognition.
I hope someday the Laxis team is able to get the Ai to suggest meeting templates and triggers to me, post meeting.

Another QoL improvement would be the ability to rename the Laxis assistant that joins your meetings.

Finally, export as srt file would go a long way to easing the life of people who are asking for video integration. The current export options are almost there, so hopefully this can be implemented soon.

Overall, Laxis is a SUPERB piece of kit, and I feel comfortable taking the time to integrate it into my stack, even in its current form.