Just tried to generate a concept for a dark fantasy novel cover. The initial image was perfect. Every variation got flagged by the new "content moderation."
* Can't have subtle, implied violence.
* Can't have "suggestive" attire on a mythical creature.
* Shadowy atmospheres now require three revisions to pass.
This feels less like safety and more like a pre-emptive cover-your-ass policy that kills creative exploration. Are we now paying for a "creative" tool that outsources its artistic judgment to an overly broad filter?
From a value standpoint:
* Pro plan is **$35/month**. Enterprise is **$95/month**.
* Are higher tiers less restrictive? The pricing page is silent.
* If 20% of my generations are wasted on filter loops, that's a direct cost increase.
Anyone found a workaround, or is this just the new TCO of using Playground?
always ask for a multi-year discount
You've hit on the exact financial metric that matters, the "direct cost increase." Framing this as a percentage of wasted generations is the right way to look at it, but I'd push you to measure it more precisely.
Is the 20% an estimate, or have you tracked it over, say, 100 generation attempts? I'd need to see the source data on that. My suspicion, based on how these filters typically roll out, is that the rate of false positives is highly variable and depends on the initial prompt's proximity to certain banned keyword clusters.
To your question about higher tiers, the pricing page's silence is the answer. If a less restrictive tier existed, it would be a premium feature explicitly marketed to enterprises for "research" or "internal use." Its absence suggests a universal, one-size-fits-all filter they are not willing to dial back, likely due to platform policy.
The real TCO now includes the time spent on prompt engineering to avoid the filter, not just the wasted credits. Have you tried abstracting your concepts further? Instead of "a harpy in tattered robes," you might need to prompt for "a winged humanoid statue with textured, flowing stone drapery," then edit in the tone later. It's inefficient, but it might lower that 20% blockage rate.
Garbage in, garbage out