Engineering
Nine Cursor switches that cut Grok Bot spend
Habits were only half the burn. Default task and cloud models were the rest.
What hurt
klöss (@kloss_xyz) already audited routines that ran too often. A follow-up: on a Cursor plan, leaving task models and Cloud Agents on default can auto-pick High/Fast and bill about 2x. He posted nine concrete settings steps and said token use dropped about 60% after.
What to install or create
No new bot. Tip only. Flip off unused models in Cursor desktop, pin a cheaper task model (do not leave Default), open Cloud Agents → Preferences → Default model and pin that too, then cap on-demand spend. Quote his steps — do not invent a prompt.
Bots
No new bot. Follow the first-run steps.
Plugins
None on day one. Follow the first-run steps.
Files
Optional: a note of which models you allow for chat vs cloud agents.
Routines
None on the first run. Watch one manual pass first.
The prompt
This is a tip, not a pasteable system prompt. The author did not publish one. Follow the first-run steps.
First run
- 1Read https://x.com/kloss_xyz/status/2095668725840244991 (full note has all nine steps).
- 2Open Cursor desktop Settings → Models. Flip off models you do not want.
- 3At the top, set Task models to a cost-efficient pick. Do not leave Default — High/Fast can bill ~2x.
- 4Open Cloud Agents (side menu → web dashboard). Under Preferences, set Default model the same way — not Default.
- 5Bonus: Spending → put on-demand usage on a cap you accept, not unlimited.
- 6Optional related tip: avoid routines every 1/5/30 minutes; prefer a daily or twice-daily schedule (see also @Bobby_1509).
Approval fence
No send, post, or spend. Changing model defaults only.
What to tweak
Steal the settings path. Default is the expensive choice twice — desktop tasks and cloud agents.
Source
klöss @kloss_xyz · Original post
Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.