This would appear to depend on some cloud service (unless it’s doing mind blowing fast lip reading or something), but the readme doesn’t have much to say about that.
The readthedocs page goes into more detail: http://subliminal.readthedocs.io/en/latest/user/how_it_works.html
It’s basically searching subtitle repositories, but also does some additional matching to find which subtitle is the best for your query.
This would appear to depend on some cloud service (unless it’s doing mind blowing fast lip reading or something), but the readme doesn’t have much to say about that.
The readthedocs page goes into more detail: http://subliminal.readthedocs.io/en/latest/user/how_it_works.html
It’s basically searching subtitle repositories, but also does some additional matching to find which subtitle is the best for your query.