{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/c36270c8630e42c2bb453fa1d1d8076f\" frameborder=\"0\" width=\"1154\" height=\"865\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":865,"width":1154,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":865,"thumbnail_width":1154,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/c36270c8630e42c2bb453fa1d1d8076f-a7ae149751e91f86.gif","duration":300.232,"title":"How Meridian Prediction Markets Work","description":"This Loom explains how the Meridian prediction market system works end to end. It covers how the Meridian program handles buy and sell orders by integrating with OpenBook V2, including the tradeoff that OpenBook v2 needs more heap stack than Phoenix Legacy but offers better interaction with the Solana program. When a market closes at 4 p.m., a scheduler hosted in Railway triggers a settle transaction, where Meridian uses a Pyth adapter price oracle to resolve the market to yes or no and update claim eligibility. An indexer service also runs in Railway to record program transactions in a SQLite database (or Postgres if scaled) so reading does not require repeated heavy RPC parsing."}