CATCH #838
DETECTED AUG 19, 2026 18:13 UTC
THE CATCHscore 6/10
EXHIBIT 838 · URGENT
Aug 19, 2026 18:13 UTC
web app
https://platform.claude.com
Anthropic's Claude platform console shows new strings for a feature flag related to detailed output token breakdowns, including internal reasoning tokens.
READY TO POST:
looks like anthropic might be rolling out more detailed breakdowns of output tokens on claude, possibly showing how much is internal reasoning.
seeing new strings in the claude console that hint at a feature flag for breaking down output tokens by category, including reasoning steps.
EVIDENCE — the receipts
⚠️ FROM WEB-APP BUNDLE — may include vendored-dependency strings, verify before posting.
Anthropic Console (https://platform.claude.com) deployed new frontend code [0c66aad6364982e4 -> 8f9cc08c319110a9]: +322 strings, -1052 strings, 3 interesting
[feature_flag] (2)
+ Breakdown of output tokens by category.
`output_tokens` remains the inclusive, authoritative total used for billing.
This object provides a read-only decomposition for observability — for example,
how many of the billed output tokens were spent on internal reasoning that may
have been summarized be
+ Number of output tokens the model generated as internal reasoning, including
the thinking-block delimiter tokens.
Reflects the raw reasoning the model produced, not the (possibly shorter)
summarized thinking text returned in the response body. Computed by
re-tokenizing the raw reasoning text, so it
[endpoint] (1)
+ /api/stripe/${i}/intent
caught by the 20-minute sweep ·
the live wire · all receipts