I love the Browser Use integration, it’s very effective at creating agentic form filling. But, I don’t understand its pricing. The cost for browser use is mostly per task, but the live output URL is from Anchor Browser, which charges per session, per hour, LLM usage, etc.
So, how are you calculating the cost of the Browser Use executions? What’s the surcharge (profit for LateNode)?
Also, why can’t I set a max runtime and/or allow the session to end? The task only takes 35s but the session keeps running until timeout anyway.
Each run has a fixed starting cost of about one cent ($0.011), and then the price per step depends on the selected model (like GPT, Claude, Gemini, etc.).
There’s a tiny markup, but it’s only to cover transaction and access maintenance costs — basically a small compensation for providing convenient access, not an extra profit.
Could you show an example of what you mean by “it doesn’t stop”? Usually, the first user action starts the browsing session, and then you need to add a small delay before calling the “get result” step using the run ID. That delay gives the session time to finish before you fetch the output - otherwise, it may stay active until it times out.
