Pay for the gate.
The agent is not the product.
The runtime that decides every tool call is the same on every plan.
The price scales with what flows through it.
What's in each plan.
Pricing questions, answered.
Any AI action that passes through the NULLRUN enforcement kernel counts as one execution: tool calls, LLM API requests, and workflow steps. A single RAG query is typically 1–2 executions (1 retrieval tool + 1 LLM call). An agent loop with 5 tool calls is 5 executions. A webhook that triggers a workflow is 1 execution.
Still have questions? Contact our sales team