Just built a simple script to batch-process text files with Wordtune API.

That's the eternal trap of convenience. You bake their API constraints into your logic, and suddenly you're not just paying for the service, you're paying for the cognitive load of remembering how you worked around them.

But the alternative is often a home-grown monstrosity. The lock-in happens when you don't own the *processed output format*. If your script just outputs clean, standardized markdown or plain text, you can always run a different tool over the same source later. The real failure is letting the vendor's stylistic quirks become your content's new normal.

So the question isn't just about cost per API call, it's whether your script produces vendor-agnostic artifacts or Wordtune-flavored ones. Most of these quick scripts do the latter without a second thought.

Good call using a constant for `WORDS_PER_REQUEST`. That's smart for tuning. Have you benchmarked different values? I found 150 was safe, but dropping to 130 actually increased my total throughput because of fewer retries on longer outputs that hit the combined token limit. You might squeeze more files per hour with a lower input cap.