Normal prompt count
Live token count from the backend worker for the untouched prompt.
---- words
Token-Cutter-5000single prompt flow
One prompt in. Two worker paths out. Live token counts the whole time.
The normal path sends your raw prompt to Grok. The TK5 path strips filler words on the edge first, then asks the same model to infer the intended meaning and answer anyway.
Backend-driven prompt metrics
Single fixed model
TK5 stop-word stripping
Prompt input
One prompt box only. Backend preview calls update the normal and TK5 token counts as you type.
TK5 prompt count
Live token count from the stripped prompt generated on the edge worker.
---- words
-- of the prompt removed by TK5 stop-word stripping
-- fewer prompt tokens before Grok sees the TK5 path
Appears after you run both workers
Appears after you run both workers
TK5 stripped prompt preview
This is the backend-generated prompt that goes to the TK5 worker path.
TK5 prompt preview will appear here after the worker returns.
Normal worker response
Raw prompt sent directly to grok-4-fast-non-reasoning.
No live run yet
Run both workers
The normal worker answer will appear here after the prompt is sent through the raw path.
TK5 worker response
Stripped prompt sent to the same model with the TK5 reconstruction instruction.
No live run yet
Run both workers
The TK5 worker answer will appear here after the stop-word-stripped path finishes.