Orchestrating Ambient Agents with Temporal | AI + Crypto Trading Demo

In this demo, Kevin Martin walks through how he built a proactive, always-on crypto trading system using Temporal for orchestration and the Model Context Protocol (MCP) for agent tools. The architecture brings together three core AI agents: Broker Agent – the user-facing entry point that manages intents and orchestrates workflows. Execution Agent – makes trading decisions and executes orders. Judge Agent – evaluates performance and continuously refines the Execution Agent’s strategy. Supporting workflows handle market data ingestion, order placement, and ledgering, all orchestrated reliably by Temporal. You’ll see how I used: ✅ Schedules to make agents proactive without custom cron jobs ✅ Signals & Queries for real-time agent-to-agent communication ✅ Temporal UI for observability and debugging ✅ Temporal Primitives as MCP Tools to make every agent action durable and auditable Inspired by ideas from the AI Engineers World’s Fair (AWS & Anthropic) and IBM’s AI Agent Meetup, this system shows how Ambient Agents can run continuously, evolve over time, and remain dependable with Temporal at the core. Code: https://github.com/Aslan11/crypto-tra... (edited) Connect with me:   / gtm-kev   --- Temporal is the simple, scalable, open source way to write and run reliable cloud applications. Learn more Blog: https://temporal.io/blog How Temporal Works: https://temporal.io/how-temporal-works Community Slack: https://temporal.io/slack Developer resources Docs: https://docs.temporal.io Courses: https://learn.temporal.io/courses Support forum: https://community.temporal.io