Topic

Engineering

Practical engineering: instrumenting LLM usage, capturing tokens across providers, and optimizing model choice and context to cut cost without cutting quality.

• Why AI guardrails belong at the edge, not in your app code — Budget checks written inside your application are easy to bypass, hard to keep consistent, and always one deploy behind. The right place to stop spend is before the request leaves.
• The four laws of a cost SDK — An SDK that sits on every AI call is in the critical path of your product. If it can break your app, it shouldn't exist. Here are the rules ours is built around.
• What actually gets captured — the honest boundary of fetch instrumentation — AtlasBurn auto-captures AI calls by patching global fetch. That's elegant and it covers most stacks — but it has a real edge, and pretending it doesn't just creates support tickets.
• Setup should be one command, not a copy-paste ritual — The gap between "installed" and "actually sending data" is where most integrations die. We closed it with a single command that shows you the diff before it touches anything.
• Model overkill — paying premium prices for commodity work — The most common AI cost leak isn't a bug. It's a premium model quietly answering questions a model a tenth of the price would handle just as well.
• Capturing token usage across OpenAI, Anthropic, and Gemini — Every provider reports usage differently, and streaming hides it in a different place again. Here's why cost instrumentation is annoying — and where to do it once.