Depends partly how we define REPL. 
And I don’t think it’s a black and white distinction. I also misspoke - meant it’s more of an issue in interactive / live programming, where it’s about having a replicable state for an end goal. Wrote something recently where I called it transactional code changes. But thinking of systems where the code at any point in time reflects the whole state of the system at that time. ie. You don’t have to replay the entire history to replicate the current state, if a resource is removed from the state it is as if it never existed, etc. More declarative less imperative.
I have a feeling you may also have addressed some of that in the framework you have?
Having written a tool for both, I’d say video is simpler. Music / audio tends to have more temporal concerns to handle - don’t tend to work with discrete frames in quite the same way.