HotBot

HotBot is an AI trading team for Solana memecoins, built for people who want to set the direction without watching charts all day. Tell it what you want to hold, set your budget and limits, and let the agents handle the day-to-day work. Our goal is to make using that trading intelligence as easy as talking to Muse. HotBot brings our trading models, Solana market data and specialist agents. It already manages themed paper portfolios with budgets, position limits and approval settings. Muse Spark makes that team easier to understand and guide through conversation. For this hackathon, we brought Muse Spark 1.3 into HotBot's portfolio interface through Meta Model API and our MCP server. We built the conversation bridge, visible read activity, expandable source records and setting proposals the user can review. The submitted video introduces HotBot's existing portfolio experience. The repository documents the new Muse conversation and links a separate walkthrough of it. The verified Muse workflow answers a real question: why did HotBot pass on ELEVEN, then buy it five minutes later? Muse compares the journal entries, explains the recorded reasons and identifies what the history cannot tell us. The user can open the exact records behind the answer. Next, the user asks to approve future buys. Muse drafts a change from Inside my limits to Ask me first and explains that ordinary sells would also need approval. The proposal stays unapplied. The trading system, portfolios and MCP backend existed before this addition. The new contribution is the Muse conversation and its source and proposal checks. This local build uses real Meta responses and HotBot paper records. The public source includes the client and new Muse layer; the existing backend remains an external dependency.