{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/261bd7da587a4bc896ab8acc7ba1b9a1\" 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/261bd7da587a4bc896ab8acc7ba1b9a1-ee8125ea215f3177.gif","duration":136.422,"title":"Substrate Shared Memory for Agent Context","description":"We built Substrate, a shared memory layer for agents that works especially well in company contexts. When feature sprints start, critical context like front end and API requirements is often missing from GitHub or Jira, which creates asymmetric information for agents. With MCP, our Cloud Code agent can write feature API context into Substrate, and another agent can read it from the shared memory layer and fetch the relevant payment schema requirements from MongoDB. We also added a visualization layer and an activity stream to monitor MCP tool calls. No action is requested from viewers."}