The Scenario

Your free-tier AI app is bleeding money. Users are sending massive 10,000-word prompts that cost $0.50 each. Your budget is $0.01 per request.

The Problem

Your code blindly sends every request to the LLM.

response = llm.generate(prompt)  # Cha-ching!

You only find out the cost after you've paid for it.

The Goal

Implement a Pre-flight Cost Check:

1. Estimate the token count of the prompt.
2. Calculate the cost (Assume $0.00005 per token).
3. If cost > budget ($0.01), reject the request.
4. Otherwise, proceed.

Requirements:

• Use len(prompt) // 4 as a rough token estimator.
• Price: $0.00005 / token.
• Budget: $0.01 / request.
• Raise ValueError if over budget.