[DOCS] Misleading messages in the Claude Code tab of Windows 11 Claude app

Documentation Type

Unclear/confusing documentation

Documentation Location

the Claude Code tab of Windows 11 Claude app

Section/Topic

the Claude Code tab of Windows 11 Claude app

Current Documentation

The Claude Code Windows app displays a message comparing total token usage to the length of Harry Potter and the Philosopher's Stone. This comparison is misleading because the book's word count represents unique, non-repeated content, while the token count in Claude Code is heavily inflated by the way the app works: every time a user sends a message, the entire conversation history is re-sent to the model as input. This means the same words are counted repeatedly within a single session, so a user's token total can reach 220,000+ tokens without them having written or read anywhere near that amount of unique text.

Because the two numbers measure fundamentally different things, the comparison does not put token usage into meaningful perspective. It is likely to confuse users who notice the mismatch rather than help them understand their usage. The message would be more useful if it either explained the re-sending mechanism, used a comparison that accounts for repetition, or was removed in favor of a clearer explanation of what token counts actually represent in this context.

What's Wrong or Missing?

The Claude Code Windows app displays a message comparing total token usage to the length of Harry Potter and the Philosopher's Stone. This comparison is misleading because the book's word count represents unique, non-repeated content, while the token count in Claude Code is heavily inflated by the way the app works: every time a user sends a message, the entire conversation history is re-sent to the model as input. This means the same words are counted repeatedly within a single session, so a user's token total can reach 220,000+ tokens without them having written or read anywhere near that amount of unique text.

Because the two numbers measure fundamentally different things, the comparison does not put token usage into meaningful perspective. It is likely to confuse users who notice the mismatch rather than help them understand their usage. The message would be more useful if it either explained the re-sending mechanism, used a comparison that accounts for repetition, or was removed in favor of a clearer explanation of what token counts actually represent in this context.

Suggested Improvement

The Claude Code Windows app displays a message comparing total token usage to the length of Harry Potter and the Philosopher's Stone. This comparison is misleading because the book's word count represents unique, non-repeated content, while the token count in Claude Code is heavily inflated by the way the app works: every time a user sends a message, the entire conversation history is re-sent to the model as input. This means the same words are counted repeatedly within a single session, so a user's token total can reach 220,000+ tokens without them having written or read anywhere near that amount of unique text.

Because the two numbers measure fundamentally different things, the comparison does not put token usage into meaningful perspective. It is likely to confuse users who notice the mismatch rather than help them understand their usage. The message would be more useful if it either explained the re-sending mechanism, used a comparison that accounts for repetition, or was removed in favor of a clearer explanation of what token counts actually represent in this context.

Impact

Medium - Makes feature difficult to understand

Additional Context

_No response_