Claude prompt token counter
See an estimated token count while writing a prompt.
Prompt templates
Reusable local prompts with variables.
Reusable variables
How it works
STEP 1
ClaudePack watches composer text locally.
STEP 2
The estimate updates after a short debounce.
STEP 3
It never claims to be Anthropic’s exact tokenizer.
Why it helps
Prompt Token Counter keeps ClaudePack focused on ordinary Claude workflows: checking status, keeping long chats understandable and reducing the small context-management chores that interrupt deep work.
Everything runs locally in the browser. ClaudePack does not add a backend, account system, ads or paid plan. When values are estimates, the interface says so. That matters for token and context information because estimates can be useful without being exact.
The feature is designed to sit inside Claude.ai without replacing Claude’s own controls. If Claude changes its interface or stops exposing a signal, ClaudePack should degrade gracefully instead of making up data.
Questions about Token Counter
Is ClaudePack's token count exact?
No. Token figures are estimates unless Claude directly exposes an exact value. ClaudePack avoids presenting estimates as Anthropic’s exact tokenizer.
Can ClaudePack estimate prompt tokens before I send a message?
Yes. ClaudePack shows an estimated token count while you write in the Claude composer.