Ok... work on (second) API copy (I call it copy for now because its a way to work on it while not turning down the main one) is going nicely even without a proper node...

HAF (hived + DB) done!
Hivemind done!
nfttracker ... started processing today...
reputation tracker ... started processing today...

I managed to understand now how to recover from aborted index creation. Not that it matters much on a fast machine (not the case here). But at least it gives me a more accurate understanding of what is happening behind the scenes.