Guides
Guides to AI cost tracking
How to see what AI costs you per customer, feature, team and tool, check it against the invoice, and bring the bill down.
LLM cost per customer: measure it, protect margin
Provider invoices never say which customer a call served. How to tag each LLM call with a customer, turn usage into margin per customer, and check the figure against your bill.
Track AI spend across providers: every AI bill in one place
AI spend arrives as API invoices, gateway reports, coding-tool usage and seat licences. How to bring them into one total, count each provider once, and reconcile metered usage against the invoice.
OpenAI cost tracking by customer, feature and team
Your OpenAI bill shows spend by project, key and model, never by customer. How to tag each OpenAI call with a customer, feature and team, reconcile the result with your OpenAI invoice, and cap the spend.
AI coding tool costs per developer, tool and team
Claude Code, Cursor, GitHub Copilot and Codex each report usage in their own way. How to collect it per developer, read cost per tool and team, check Claude seat use, and set the spend next to merged pull requests.
Reduce LLM API costs: routing, caching and batch pricing
Find which actions, features and customers drive your LLM bill, then lower it with cheaper-model routing, response caching, batch pricing, fewer tokens and spend limits, and measure the saving honestly.
Anthropic API cost tracking: Claude cost per customer
Your Anthropic bill shows what you spent, not which customer it served. How to tag each Claude API call with a customer, feature and team, check the figures against Anthropic's cost report, and cap the spend.
AWS Bedrock cost tracking by customer and team
Track Amazon Bedrock costs per customer, feature and team from your code, including on AWS Lambda, check them against Cost Explorer, and avoid the inputTokens cache mistake.
Gemini API and Vertex AI cost tracking per customer
Track Gemini API and Vertex AI costs per customer, feature and team, see which SDK and router paths cover each, and check the figures against your Google Cloud Billing export.
Azure OpenAI cost tracking by customer and team
Your Azure bill shows the total, not which customer a call served. How to tag each Azure OpenAI call by customer, feature and team, price deployment names, and check it against Azure Cost Management.
Vercel AI Gateway cost tracking by customer
Vercel AI Gateway's spend report shows models and tags, never which customer a call served. How to tag each Gateway call, bring in Vercel's spend report, line up its tags and cap the spend.
OpenRouter cost tracking by customer and team
OpenRouter reports what your account spent, never which customer it served. How to tag each OpenRouter call with a customer, feature and team, check it against OpenRouter's account report, and get spend alerts.
Token Meter vs OpenRouter vs LiteLLM: savings and fees
How Token Meter, OpenRouter and LiteLLM measure what routing and caching saved, what each counts when a price is unknown, and what each charges, our higher credit fee included.
Setting up Token Meter itself? Start with the quickstart, browse the docs, or compare plans.