I have been doing another round of practical work on my two Premiere subtitle projects.
The older one lives at mogrt.zinchuk.online and grew out of the text-to-tiktok codebase. Its tool page is here: Adobe Premiere Text-to-MOGRT.
The newer one is Subtitle Lab, which handles a broader roundtrip workflow between MOGRT, SRT, and rewritten Premiere projects. Its public description lives here: Subtitle Lab.
This latest update was less about flashy new features and more about making both tools behave better on real editor-made projects.
What changed
The main improvement is structural.
Both projects now rely less on overly narrow assumptions about where a MOGRT keeps its useful parameters. Instead of trusting only the first component or the first obvious parameter, the newer logic walks through the full component chain and inspects more of the project structure before making decisions.
That sounds technical because it is technical, but the user-facing effect is simple: the tools are more likely to survive messy real projects.
I also added stronger preflight checks around template tracks and clearer diagnostics for newer Premiere project versions. In practice that means the services can warn earlier, fail more honestly, and waste less time pretending a broken setup is fine.
Subtitle Lab got the most visible polish
On the public side, Subtitle Lab now does a better job of showing important warnings immediately after project inspection instead of hiding them lower in the workflow.
I also corrected one subtle but important behavior in the MOGRT -> MOGRT scenario.
The service should keep motion and styling from the template track, not accidentally borrow those runtime parameters from the source subtitles. That is now fixed, which makes the result much more consistent with the whole point of the workflow: restyle the subtitles without inheriting the wrong movement behavior.
Why this matters
This kind of update is easy to underestimate because it does not announce itself with a dramatic new button.
But these are the changes that make a tool feel trustworthy.
If a service can inspect the wrong kind of project more intelligently, warn you sooner, and preserve the right template behavior when rewriting subtitle clips, it becomes much easier to use without second-guessing every export.
That is especially important now that the second project is no longer just an experiment. Subtitle Lab already has three real workflows, and each of them benefits from stricter assumptions, better inspection, and less magical behavior under the hood.
The nice part
One thing I like about this phase is that the second project is now feeding lessons back into the first one.
The old service helped me understand Premiere project structure in the first place. The newer one is now forcing cleaner abstractions, better diagnostics, and safer behavior. That back-and-forth is exactly what I hoped would happen once both tools became real enough to test against different kinds of projects.
It is still the same larger theme: subtitle work in Premiere should require less manual recovery, less guesswork, and fewer fragile one-off fixes.
That is still the direction.
If you want to see the two tools in their current public form, they are here: