// the blog
Keeping AI costs under control
Field notes on LLM cost control - spend caps, per-user limits, spike alerts, and monitoring Claude and GPT bills before the invoice lands.
- ControlsJul 18, 2026 · 6 min
Per-User Rate Limiting for LLM Features
Per-user LLM rate limiting protects both cost and availability. How to bound bursts per user across Claude and GPT, and how it differs from a spend cap.
- Cost spikesJul 17, 2026 · 7 min
Anatomy of a $10k LLM Bill: A Post-Mortem
A blow-by-blow post-mortem of an LLM bill spike: how a small change turned a $300 month into a $10k one, why nobody noticed for three weeks, and the controls that would have caught it.
- ControlsJul 15, 2026 · 6 min
Setting max_tokens Correctly (and Why It's a Cost Control)
max_tokens is the simplest cost control you have and the most commonly misused. How to set max_tokens correctly across Claude and GPT to cap output cost without truncating real answers.
- MonitoringJul 14, 2026 · 5 min
Sending LLM Cost Alerts to Slack
An alert nobody sees is decoration. How to route LLM cost anomalies to Slack with enough context to act - which feature, which user, how far from baseline - without alert fatigue.
- MonitoringJul 11, 2026 · 7 min
Monitoring Anthropic Claude API Costs in Production
How to monitor Anthropic Claude API cost in real time: attribute spend by feature and user, track tokens per call, and catch problems before the monthly invoice.
- Cost spikesJul 4, 2026 · 6 min
Why AI Agent Costs Compound Instead of Add
AI agent cost doesn't grow linearly with traffic - it grows with traffic times loop depth. Here's why agentic workloads compound, and how to bound them.
- ControlsJun 27, 2026 · 6 min
How to Set a Hard Spend Cap Per User on LLM APIs
Enforced budgets, not advisory ones. How to implement a hard spend cap per user across Claude and GPT so one person can't drain your whole AI budget.
- Cost spikesJun 18, 2026 · 7 min
How to Stop Your OpenAI API Bill From Spiking
A step-by-step playbook to prevent a surprise OpenAI or Anthropic bill: find the spend, cap the worst offenders, and get alerted before the invoice - not after.
- FundamentalsJun 9, 2026 · 8 min
LLM Cost Control: A Practical Guide for Production Apps
A field guide to LLM cost control: where AI API spend actually comes from, the four controls that keep it predictable, and how to enforce budgets without slowing your app down.