I saw the same problem I have encountered with developer tools ever since: each product can document its own controls, but developers still need someone to show them the path across product boundaries—from a text source to a character who speaks, moves, tracks, and looks believable enough to test.
Why I made this
This was not a one-plugin setup. A developer had to acquire the character assets, install several Unity packages, configure a text-to-speech speaker, connect SALSA to the MCS morph system, add eye tracking and animation, and prevent the avatar jaw from fighting the lip-sync behavior.
That integration surface created several likely failure points. Even when every product worked individually, developers could still get stuck on package order, component relationships, animator state, morph configuration, or a final result that technically ran but did not look correct.
What I did
I began the video description with a concrete download checklist: MCS and its Unity assets, SALSA and its MCS helper, RT-Voice, and a Mixamo idle animation. That made the dependency surface visible before the viewer entered Unity.
Inside the project, I followed the integration order rather than explaining each product in isolation. I set up the character, configured RT-Voice and its text-file speaker, attached the SALSA helper, and ran an early test before adding more behavior.
Once the speech path worked, I tuned the speaking controls and eye tracking, imported an idle or talking animation, added the controller, and tested again. I finished by adjusting the avatar jaw so the animation rig and lip-sync system did not compete for the same movement.
The tutorial attracted detailed implementation questions and positive feedback, including recognition from the SALSA developer. More importantly, those questions exposed where developers still encountered friction—one-click setup errors, animator state changes, jaw handling, and dynamic emotion—which is useful product and documentation feedback.
A note on the archive
This recording was published in April 2017 and is explicitly deprecated. The original MCS, SALSA helper, RT-Voice, asset-store, and Mixamo workflow should not be assumed to remain available or compatible with a current Unity project.
There were no usable captions on the source video, so I didn’t invent chapter timestamps. Use current vendor documentation and supported replacements if you want to build something similar. I’m keeping the page because it shows the entire integration problem, including where one product handed off to the next.