How to beginner · 3 min read

AI workflow cost monitoring

Quick answer

Use API usage endpoints or billing dashboards from providers like OpenAI or Anthropic to programmatically track your AI workflow costs. Combine usage data with your pricing plan to calculate expenses and monitor spending in real time.

PREREQUISITES

Python 3.8+
API key for your AI provider (e.g., OpenAI or Anthropic)
pip install openai>=1.0 or anthropic>=0.20

Setup

Install the official SDK for your AI provider and set your API key as an environment variable. For example, with OpenAI:

Run pip install openai
Set OPENAI_API_KEY in your environment

bash

pip install openai

output

Collecting openai
  Downloading openai-1.x.x-py3-none-any.whl (xx kB)
Installing collected packages: openai
Successfully installed openai-1.x.x

Step by step

This example shows how to fetch usage data from the OpenAI API and calculate estimated costs based on your pricing plan.

python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Fetch usage data for a billing period
usage = client.billing.usage.list(start_date="2026-04-01", end_date="2026-04-30")

# Example pricing (USD per 1K tokens) - update with your plan
pricing = {
    "gpt-4o": 0.03,  # $0.03 per 1K tokens
    "gpt-4o-mini": 0.015
}

# Calculate total cost
cost = 0.0
for record in usage.data:
    model = record.model
    tokens = record.total_tokens
    rate = pricing.get(model, 0)
    cost += (tokens / 1000) * rate

print(f"Total tokens used: {sum(r.total_tokens for r in usage.data)}")
print(f"Estimated cost: ${cost:.2f}")

output

Total tokens used: 125000
Estimated cost: $3.75

Common variations

You can adapt this approach for other providers like Anthropic by using their billing or usage endpoints. For asynchronous workflows, use async SDK methods if available. For streaming usage data, poll periodically and aggregate costs.

python

import os
from anthropic import Anthropic

client = Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])

# Hypothetical usage fetch (replace with actual API call if available)
usage_data = client.billing.usage.list(start_date="2026-04-01", end_date="2026-04-30")

# Process usage_data similarly to calculate costs
# Note: Anthropic pricing and API may differ; check docs.

output

Usage data fetched successfully
Estimated cost: $4.20

Troubleshooting

If you get authentication errors, verify your API key is set correctly in the environment.
If usage data is empty, check the date range and ensure your account has usage in that period.
For unexpected cost calculations, confirm your pricing rates match your provider's current plan.

✅

Key Takeaways

Use official SDK billing or usage endpoints to programmatically monitor AI costs.
Combine usage data with your pricing plan to calculate accurate expenses.
Automate periodic cost checks to avoid unexpected billing surprises.

Verified 2026-04 · gpt-4o, gpt-4o-mini, claude-3-5-sonnet-20241022

Verify ↗