# how do you separate Claude API spending between staging and production? — retained record

> Retained reddit record from reddit:ClaudeAI.

Canonical: https://highsignal.app/data/records/a1d3e610244c4ba0

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeAI)  published 2026-10-31 15:05 ingested 2026-11-01 02:36

[original source ↗](https://www.reddit.com/r/ClaudeAI/comments/1xkcjql/how_do_you_separate_claude_api_spending_between/)

## retained text



```
We've been using the Claude API for a few different features and lately I've been having trouble figuring out how much of our spending is coming from staging versus actual production usage. Our staging environment gets used pretty heavily for testing prompts/trying out changes before deploying them and sometimes we'll run the same requests multiple times to compare outputs so the token usage adds up even though none of it is coming from real users.

As of right now we can see our overall Claude API costs but I'd like a clearer breakdown of what we're spending on testing versus production without having to manually go through usage logs so if there are any teams running Claude in both environments, are you using separate API keys or projects to track spending or have you found a better way to manage it?
```



document fetched 2026-11-01 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xkcjql/how_do_you_separate_claude_api_spending_between
