Q: How much does each “variable run” cost in terms of AI credits?
I understand that there are fixed costs to first set up the AI agents workflow. (2-4 credits for each sub-step).
I also understand that every time we ask it to generate/modify a AI Agent code, it costs credit. Question 1: is it also 2-4 credits per generation?
Question 2: when a user uses my Webapp once my webapp goes live, does it cost me credit every time it processes the user’s request? If so, is it also 2-4 credits per sub-step per generation?
Can you please elaborate more with detailed examples?
SeanP_AgenticFlowAI
May 8, 2025A: Hey Slock08,
Gotcha. Think of it this way: every action your workflow or agent takes costs a small fixed amount of your AgenticFlow credits just to run on our platform, like 3 or 4 credits per step depending on your tier.
Now, if a step uses one of our built-in AI models like Gemini Flash or 4o-mini to generate something, that specific step costs extra variable credits based on how much thinking (tokens) it did.
But, if you plug in your own API key for a model like GPT-4 or Claude (BYOK), you don't pay us the variable credits for that AI thinking part – you pay OpenAI or Anthropic directly. You still pay us the small fixed credit cost for running that step, though.
And yes, absolutely – every time an end-user triggers your live agent or workflow, it runs those steps, and those steps consume credits according to that same fixed + (maybe) variable logic. Hope that simplifies it!
Slock08PLUS
Verified purchaser
Thanks Sean. So, the "variable credits" isn't 3-4 credits per generation, but can go higher? Do we have an idea of the range (for example, a simple step like "create a form, with this format" would take 10 credits)?
And everytime an end-user triggers a workflow/agent, how much credits will it consume (for example, if it takes 10 steps to return a result per query, then 20 credits/step x 10 steps?