來自 @fedeecaccia 的 Grok Bot 案例

嘿！我今天的建議：別叫 Grok Bot Agent 直接執行程式碼。改叫它們去編排雲端的 Cursor Agent。 為什麼？ - 編排層留在 Grok Bot。 - 你可以在 @cursor_ai 選不同模型和多個 Agent。 - 最重要的是，你不會燒光寶貴的 Grok Bot 餘額。執行改走你的 Cursor 餘額。底層計費相同，但兩邊餘額消耗速度差很多。 這套設定對我很好用。 你在專案裡是怎麼編排 AI Agent stack 的？

原文：Hey! My recommendation of the day: don’t ask Grok Bot agents to execute code directly. Ask them to orchestrate Cursor agents in the cloud instead. Why? - You keep the orchestration layer in Grok Bot. - You can choose different models and multiple agents in @cursor_ai . - Most importantly, you don’t burn through your valuable Grok Bot balance. The execution happens through your Cursor balance instead. Same underlying billing, but the two balances are consumed at very different rates. This setup has been working really well for me. How are you orchestrating your AI agent stack in your projects?

整合