The animations, sound and concept are very good although for now the gameplay seems weak to me due to its early state and lack of tension, difficulty, meaningful goals and progression. I think more real-time gameplay outside of the struggling could be nice, for example combat, platforming or stealth (the gameplay could really go in any direction since there isn't much there yet). The sandbox would feel more believable if enemies pursued their goal (presumably abduction of the player character) with more speed and aggression and if their actions had more consequences (i.e. abduction, loss of items or resources or otherwise).
I think one problem for me is that enemies don't really feel like enemies because they stand and walk around aimlessly rather than actively looking for you, but they also don't feel like civilians because they pursue you when they see you (but then do so very slowly and give up easily). What are they doing there in the first place? For example, what if instead there were actual enemies and actual civilians and the enemies ("cops"), which are more challenging to capture and an actual threat, protect the civilians, which are easier to capture but have to be captured without getting caught by patrols. This type of scenario, which is just a quick example, could create a more believable and dynamic sandbox and create an actual gameplay experience. But it's just one idea to point out the kind of thing I see missing right now. We could fulton out captives instead of having to leave the area so as to maintain game flow.
I agree with the previous poster that the struggle minigames are too visually distracting and take up too much visual space.
If you don't want just a simple "A-D" spam minigame, for a struggle system in this kind of game with 3D models and mouse control, I think it might be more interesting if, for example, we could mouse-select body areas on the character being subdued (whether enemy or player) to "attack" and struggle against, and then weaken them with a minigame until perhaps an overall escape/subdue meter is filled or a certain number of areas are "broken". This could keep the visual attention on the animation, and make the struggle system feel more involved, visceral and immediate.
I also think that other kinds of minigames might be viable in that kind of system, for examples ones that require precise timing or something akin to the various fishing minigames that have been done over the decades.
If you have a purely skill-based struggle system without randomization, I think the game needs to have systems in place to create unpredictable and intense situations that cannot be overcome by simply mastering minigames. That could be stamina (i.e. a character getting caught after running for a while will be at a disadvantage); combat, traps and ranged attacks (a weakened character is at a disadvantage); stealth, arousal, state dependence (a surprised, bound or aroused character is at a disadvantage), escalating difficulty or stat differentials (some enemies being significantly more difficult and powerful, a leveling system) which weaken a character's struggle power in any given situation.
To prevent the opposite situation, where struggling in certain situations is already doomed from the get-go, you could have various usable items that enhance your odds but have to be used strategically.
