{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/a5f74aedf07e4727afaedbd77741ee0b\" frameborder=\"0\" width=\"1662\" height=\"1246\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1246,"width":1662,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1246,"thumbnail_width":1662,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/a5f74aedf07e4727afaedbd77741ee0b-44497e398485ea6d.gif","duration":524.216,"title":"Building a Crypto AI Coding Terminal","description":"This Loom demos PF Terminal, a Codex fork that runs AI agents and models entirely in crypto. The speaker shows topping up a Solana wallet with about 20 to 21 USDC, purchasing a $20 PF Terminal plan, and then generating a landing page using GLM 5.2 while highlighting multi-model support across many providers. They explain that PF Terminal runs on its own GPUs, supports models like GLM 5.2 or DeepSeek Flash, and can run larger ZAI setups on eight GPUs. The Loom also contrasts performance with other tools like Hermes and Cloud Code, noting that those load expensive instruction blobs into prompts, which can slow things down and conflict with user directions."}