Okay, I just *have* to get this off my chest and see if I'm running into a unique issue or if this is the shared experience. I've been deep-diving into Firefly for the last few months, integrating it into our analytics team's workflow for generating quick visuals and mockups for dashboards. The quality for certain things is fantastic, especially the vector recoloring and text-to-template stuff.
But good grief, the credit burn rate is *astronomical*. I feel like I'm watching the meter spin every time I hit "generate." Here's my typical session:
* **Iterative design:** Let's say I need a custom icon for a "data pipeline" concept. I start with a prompt, generate 4 variations. That's 4 credits. Maybe one is close, so I use "Generate Similar" a couple of times. Another 4-8 credits. Tweak the prompt, generate 4 more. It adds up so fast you barely notice until you've blown through 30 credits on *one icon*.
* **The "High-Quality" toggle:** I get that this costs more, but turning it on feels almost mandatory for anything going to a client. That instantly doubles your consumption. A single 4-image set at high-quality is 8 credits. Ouch.
* **Compared to other services:** I also dabble in Midjourney and DALL-E. While they have their own pricing, the "per credit" anxiety doesn't feel as acute. With Firefly, because it's baked into the Creative Cloud ecosystem I already pay for, I guess I expected the "allowance" to feel more generous for the tier. My 100 monthly Premium credits vanish in a few focused sessions, and then I'm stuck either buying a top-up (which stings) or waiting for the refresh.
My main gripe is that the creative process is inherently about iteration. You rarely nail it on the first shot, especially with more niche, specific concepts (like, say, a "self-documenting ETL process visualized as a garden" – yes, I've tried that 😐). The current credit cost per generation actively discourages that exploration, which is half the fun and utility!
Is anyone else feeling this pinch, especially those using it for more than just casual experimentation? Have you found any workarounds or specific workflows that help conserve credits? Maybe I'm just using it "wrong." Would love to compare notes.
—Jake
Spreadsheets > opinions
You are definitely not the only one. Your breakdown of the iterative design process mirrors my own documented experience with the platform, particularly for compliance-approved assets where we cannot allow any stray elements. The high-quality toggle isn't just "almost mandatory" for client work in my field, it is an absolute baseline requirement, which functionally halves the practical value of any credit bundle.
What I've started doing is maintaining a separate log of prompt iterations and the resulting credit cost per finalized asset. This creates a clear metric for cost-per-output. When you compare it to a fixed-cost stock asset purchase, the tipping point becomes very visible. For one-off, highly specific items, Firefly can still be justified. For a series of icons or standardized templates, the cumulative credit consumption you described often exceeds the cost of commissioning or buying a bulk pack.
Have you looked at the usage analytics in your account dashboard? The monthly report can be sobering. It shows the concentration of credits in specific feature areas, like text-to-image versus vector recoloring. That data might help you build a case for a more tailored subscription, if one exists, or force a stricter pre-generation prompt refinement phase.
Exactly. That iterative cycle you described is the hidden cost center they don't highlight in the feature list. The "Generate Similar" button is a credit furnace, especially when you're chasing a specific, coherent style. It feels less like a tool and more like a slot machine where you pay per pull. You start rationalizing, "just one more set and I'll have it," and suddenly you're out 30 credits and still not satisfied. At that point, commissioning a human designer looks less like an extravagance and more like prudent financial planning.
You touched on the comparison to other services but trailed off. I'd be curious to see that breakdown, because my experience is that while Firefly burns faster, the per-credit output is often more usable off the bat, ironically justifying the burn in some twisted way. The real question is whether that initial quality advantage evaporates once you factor in the iterative tweaking to make it *exactly* what you need.
The high-quality toggle is just insult to injury. It creates a two-tier system where anything for professional use is automatically taxed at a 100% premium, making the advertised credit bundles a complete fiction for actual work.
Trust but verify.