Currently the sampleapps/messenger-platform/ sample does some ugly code here to find a translator by hard coded path then call methods directly on it.

Instead, we need to use the opent2t library and do this correctly, similar to how the cloud sample does it here.

We also should not have to download the entire git repo as we do in git.js and towards the end of app.js. All that should be cleaned up to use the official methods in the opent2t library.

Currently the sampleapps/botframework/ sample is just parsing commands and not actually invoking any translator. We need to add code approximately here to use the opent2t library, similar to how the cloud sample does it here.

We also should not have to download the entire git repo as we do in git.js and towards the end of app.js. All that should be cleaned up to use the official methods in the opent2t library.

sampleapps/alexa/ starts an alexa sample, specifically the speech assets.

It should be really simple to create a skill that implements "Alexa, turn on the lights" via opent2t. See the sampleapps/messenger-platform/ sample for motivation.

Also, note you can just use echosim.io to develop/test this sample. You don't really need a real echo.

Is there any way I can discover devices(which may be different manufacturer)? Best case scenario I give a router/hub address to openT2T and it tells me that there are 5 connected devices. 2 of them are from manufacturer X(among them one a light, another is micro oven) and 3 are from Y.