{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/65b10da21d9140f3bf6f3930d40aabd0\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/65b10da21d9140f3bf6f3930d40aabd0-49c3157a984359e6.gif","duration":5651.4,"title":"Claude Token and Prompt Optimization Cheats","description":"This Loom explains behind the scenes Claude and AI “token cheat codes” for reducing cost and improving output quality. It covers cleaning up global and project instructions, turning web search on only when needed, using extended thinking for big decisions, managing memory (including updating or disabling “generate memory from chat history”) and deleting irrelevant project chats monthly or quarterly to avoid context overload. The speaker also shares practical token-saving tactics like batch instructions in one message, restarting or editing instead of continuing long threads, using the Claude token counter extension, and trimming files while preferring Markdown over PDFs for fewer tokens. It further demonstrates automation via Chrome extensions, skills, connectors, and tasks, including scraping Instagram reels and running it on schedules."}