Skip to content

Pricing

Pay for tokens, not for seats.

One credit is one cent of what the model provider charges us. A run on Gemini Flash Lite costs a fraction of a credit; a run on Opus costs a few. You spend what you actually use — or bring your own key and spend nothing here at all.

Free

$0/mo

Enough to decide whether the frames are worth it.

30 credits / month

  • 30 credits a month
  • Gemini Flash Lite and Claude Haiku
  • The full vault

Pro

$12/mo

The working tier. Every mid-weight model, room to actually iterate.

400 credits / month

  • 400 credits a month
  • Adds Sonnet 5, Grok 4.6, Kimi K3, Gemini 3.5 Flash
  • Optimization history and compare runs kept

Power

$29/mo

The frontier models, for prompts that are going into production.

1,000 credits / month

  • 1,000 credits a month
  • Adds Claude Opus 5 and GPT-6 Astra
  • Priority support

BYOK

$8/mo

You bring the keys. We never touch your credits or ours.

Unlimited on your key

  • Unlimited runs on your own API key
  • Every model, including Claude Fable 5.1
  • One OpenRouter key covers all of them

Credit top-up

A one-time 300 credits, on whatever models your plan already unlocks.

$10+300 credits

What a run actually costs

Per model, per optimizer run

Estimated on a typical run of 1,500 input and 1,000 output tokens. You are charged on the real token counts the provider reports, so a short prompt costs less than this and a long one costs more.

ModelCredits / runRuns on 400 creditsSmallest plan
Gemini 3.1 Flash Lite0.192,133Free
Claude Haiku 4.50.65615Free
Gemini 3.5 Flash0.56711Pro
Claude Sonnet 51.30307Pro
Grok 4.61.80222Pro
Kimi K31.95205Pro
Claude Opus 53.25123Power
GPT-6 Astra14.0028Power
Claude Fable 5.11.30307BYOK