Credits System

Deep dive into how credits are consumed, estimated, and optimized.

How credits work

Credits are consumed when an AI plugin executes. The cost varies by model — more powerful models cost more credits. Non-AI nodes (logic, input, publishing) are always free.

Credit consumption

CategoryPluginCredits
AI TextChatGPT (GPT-4o)2
Claude (Sonnet)3
Gemini (Flash)1
AI ImageNanoBanana (Flash)3
DALL-E 35-15
AI VideoVeo 3.1 Fast20
Sora 215
AI AudioElevenLabs Flash2
Logic/Input/PubAll0

Credit estimation

Before executing, the editor estimates total credits based on the nodes in your workflow. The estimate appears next to the Execute button.

Estimates vs actuals: Text generation costs may vary slightly based on output length. Image and video costs are fixed per execution.

Optimization tips

  • Use Gemini Flash over Pro for 50% savings on text (1 vs 2 credits)
  • Use NanoBanana Flash over DALL-E for images (3 vs 5-15 credits)
  • Use Sora 2 over Sora 2 Pro for video (15 vs 30 credits)
  • Gate expensive nodes with If/Else to avoid unnecessary AI calls
  • Use EdgeTTS (0-1 credits) instead of ElevenLabs (2-5) for prototyping

Credit rollover

Unused credits do not roll over to the next month on any plan. When your monthly allocation refreshes, your balance resets to your plan's credit amount.

Last updated: 2026-03-27

    Credits System | Zephly