claudoscope

Preset
Model

Any model id your key supports — type your own.

System prompt

4,592 chars · caching needs ~1k+ tokens to engage

Messages
Max tokens · 512
Temperature · 1.0
idle
no response yet

Run a request to stream a live response and x-ray its token economics.

X-Raytoken economics of the last run
Token composition0 tokens
run a request to populate
cache read0
cache write0
uncached input0
output0
Cost
$0
est · sonnet pricing
TTFB
time to first token
Total time
request to finish
Speed
output throughput
Cache impact

No cache activity. Turn on the system-prompt cache and give it ~1k+ tokens, then run twice — the second run reads from cache at a ~90% discount.