The plans, in one table
| Plan | Price | Prompts per month | Notable differences |
|---|---|---|---|
| Starter | $9 / month | 100 | Shared generation pool, standard quality, commercial licence |
| Pro | $19 / month | 200 | Faster queue, extra-high thinking for complex scenes, 2 parallel sub-agents, higher quality mesh output |
| Premium | $49 / month | 800 | Priority speed, best mesh quality, 4 sub-agents, batch scene generation, longest context |
Annual billing saves 17% on the same allowances: $90, $190 and $490 respectively. See the full plan comparison.
How prompt allowances work
One prompt equals one generation request. Three rules matter in practice:
- The allowance resets monthly. Unused prompts do not roll over, so a plan is a per-month budget rather than a bank.
- Failed generations should not cost you. A request that fails for a technical reason is returned to your balance instead of being consumed.
- Vague prompts waste the same allowance as good ones. A brief with object, style, scale and destination lands closer to the target on the first attempt, which is the cheapest optimisation available on any plan.
What "free tools" cost instead
The alternatives to a paid assistant are not zero-cost - they move the cost somewhere else:
- Local open-source pipelines cost GPU time, setup effort and maintenance every time a model or dependency changes.
- Web text-to-3D generators often include a small free allowance, then meter credits; their output still needs an import-and-repair pass in Blender.
- Doing it by hand costs hours, which is the most expensive option for repetitive assets like props, kit pieces and set dressing.
If you generate two or three assets a month, a free web tool plus a cleanup pass may genuinely be enough. If you generate assets every week, the comparison usually flips - see the alternatives breakdown.
Where the money goes, plainly
Starter at $9 for 100 prompts is roughly nine cents per generation before any cleanup time. That is the number to compare against your own hourly rate: a prop that takes an hour to block out by hand versus a minute of generation plus ten minutes of cleanup.
Premium at $49 for 800 prompts exists for continuous production - teams generating through a whole level, or iterating on the same scene daily. It is not a better tool; it is the same tool with a bigger budget and a shorter queue.
