Token and Time Counter Missing in Normal Mode 1.0.83

Found 1 possible duplicate issue:

1. https://github.com/anthropics/claude-code/issues/5929

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

You can add it using /statusline command

Status Line Configuration

You can add it using /statusline command Status Line Configuration

@marcoscale98 do you mind sharing how? The JSON data that's available in the input structure doesn't contain anything about the token counts so I'm not sure how to reproduce it using the statusline feature.

> You can add it using /statusline command > > Status Line Configuration @marcoscale98 do you mind sharing how? The JSON data that's available in the input structure doesn't contain anything about the token counts so I'm not sure how to reproduce it using the statusline feature.

I have tried it, but I noticed that it's bugged, maybe because it's not have data from JSON input

+1 we need this back

100% this has to be coming back, otherwise it is not possible to tell if CC is just stuck (which does happen, especially behind various wonky internal corporate proxies) or legitimately streaming e.g. multi-thousand token response.

Was pretty confused to see this suddenly disappear.. can we please have this back?

It’d be great if we could get the token counter back in normal mode.

• Super useful for keeping an eye on context size, debugging prompts, and tracking cost.
• Right now the only option is --verbose, which works but clutters the output a lot.
• The simple counter was one of the most handy features, and losing it really slows things down.

Any chance it could be restored, or maybe just a flag like --show-tokens so we can toggle it without full verbose mode?
Would make life a lot easier for those of us using Claude Code every day.

Agreed, --verbose is not a suitable solution, it is not a mode you would like to run day to day (too much noise), perhaps useful for debugging, etc.

On the other hand, token counter is immensely useful - especially when dealing with flaky proxies, providers, etc.
PLEASE add option to bring it back, perhaps via statusline parameter or something like that.

Claude Code team has added the /context command recently that comes close to what you want.
I think It won't be long before we see the tokens in the statusline as well.

For Time Counter, it has been added the cost.total_duration_ms in the JSON input for the statusline. You can configure it via /statusline command

people mentioning statusline or context command are giving incorrect information. This token count is specifically required during the streaming of cc , and helps to work out if cc is hung or streaming a large response. This is different to the token usage percentage stats relevant to the entire cc session. Right now there is no way to configure this feature unless verbose mode is on where most people don't need the full debug logs.

+1, what version can we npm install -g downgrade to to get those back?

~~this fixes it, in fact I think this can probably close the ticket:
/config -> enable verbose mode~~

correction: while this works, verbose is far too verbose! there should be a setting just for the token ticker:

<img width="483" height="118" alt="Image" src="https://github.com/user-attachments/assets/1274e9a3-3e17-4645-8c00-ac95b8a5fd64" />

Hi everyone! We just added a feature to tweakcc to re-enable the token counter and the elapsed time automatically without having to run Claude Code with --verbose option. To try it out, just run npx tweakcc --apply . Enjoy!

Bring it back please

+1 Please bring it back! I'm running out of limits without ever knowing what's happening and how I can stop it when I'm working under pressure.

The users want it! I've cancelled one claude subscription because of the new limits, and if I could see the tokens I'm using and use that information to properly tune my configuration, maybe that will avoid it happening with others or even cause me to give you more money!

ccstatusline allows you to configure the status line to see context window token count, input token count, output token count and cache token count

Please bring this back

Yeeaaah I'm gonna call conspiracy here. This is a fucking psyop. I don't know any self-respecting developer in existence, an engineer of AI at the frontier no less, employed at Anthropic, who would do such a thing given the regime involved here. We all know API requests can drop, we've all seen it. You need a pulse on the connection. This is a dark pattern in business where you obfuscate the numbers and create unnecessary fog of war, this way your customers can't track when the regime has changed in a way that doesn't benefit you, uses more tokens for the same functionality. Can't tell when the ingredients changed, shaved off a couple grams of product weight. It gives them pathways for frog boiling, slowly reversing quality and normalizing it, whereas with a precise pulse on every number the users are 24/7 keeping a tab psychically, performing free energy minimization in their brain to drive the token count lower and lower for the same behavior or better. This is all aligned with immoral business practices as well as their anti-acceleration constitution.

Perhaps the official removal of this feature is due to their lack of confidence in the claude code official app speed, as the actual rendering would reveal their poor turtle speed.

The speed of recent claude code is too slow.

I'm on v1.0.111 and I can see the token count and timer is back!

<img width="1104" height="190" alt="Image" src="https://github.com/user-attachments/assets/c4da02b4-5f0d-4c95-bd27-f2c07c993bd5" />

@coygeek I also on v1.0.111 but can't see that, you enable some flags?

<img width="1098" height="654" alt="Image" src="https://github.com/user-attachments/assets/f53e50ba-1921-4bae-86db-283ddcf3c452" />

<img width="1098" height="654" alt="Image" src="https://github.com/user-attachments/assets/20443898-5336-4fbf-a3af-f90f1488fddc" />

Nope.

Looks like I'm not the only one: https://www.reddit.com/r/ClaudeAI/comments/1ndufe0/so_they_finally_restored_token_usage_so_that_we/

Cool background!

@coygeek I can see that now, seems that will appear only when real response and slow

@oxysoft

you wrote it quite a weird way, but you're right, they tried to obfuscate internal information, however if someone is really dedicated they could get this information through side channels anyway, so they figured out it's not worth to hide token count and probably got enough complaints.

This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.