3D printed key mechanism for a celesta

I am building a keyboard instrument with a glockenspiel suspended above a set of 3D printed key mechanisms, which makes it really a celesta rather than a piano. Each note's whole implementation is narrow enough to fit within the width of a key, and alongside the key mechanism I have prototyped a sustain damper that lifts off the tone bar when the key is pressed and returns to damp it when the key is released. After trying to make the mechanisms compliant, I have given up on that approach and I am going to make an ordinary mechanism instead.

The plan is to suspend a sort of glockenspiel above a bunch of these key mechanisms so that it is a working "piano", though actually it is more of a celesta. I figure nobody has heard of a celesta, so I tend to say piano. A harpsichord has strings and the key mechanism plucks them, whereas a celesta has literally like a glockenspiel inside. Here is an example of one, although mine will be super basic; I'm not a real musician.

With my printed key mechanism I am planning to make the entire implementation of each note narrow enough that it fits within the width of a key, which makes the whole thing much simpler, because then you don't need multiple banks of glockenspiels.

I did some experiments with trying to make a keyed string instrument before, and it is actually really hard to get a decent sound out of a string, so I gave up on that. But if I just cut some lengths of steel bar to make a glockenspiel and hit it with plastic hammers, it is bound to work. What could go wrong?

This photo shows two mechanisms: the one in the front is the key mechanism as before, and the one in the back is a "sustain damper" mechanism, so when you press the key the little hammer moves down and stops contacting the bar, letting the tone sound, and when you release the key it contacts the bar again to damp the sound. It pretty much looks like it will work, but I need to attach a tone bar in position to check for sure; so far I am just holding the tone bar above it by hand. Video of it working here. This solves a different problem from repeated hits; it is just an extra feature that gives you more control over how long the note is sustained, and I am ignoring the repeated hits problem until I am sure it is a real problem.

Latest video. Compliant mechanisms are too much trouble, so I'm now officially giving up on it and I'm going to make an ordinary mechanism instead. If I had done that in the first place, then the extra time I spend assembling 30 copies of it would have more than paid for itself in the time I didn't spend messing about trying to make compliant mechanisms work.