If you're building AI trading agents, this solves the memory problem. Your agent can execute trades but can't remember what happened last time it saw similar conditions. TradeMemory gives you 17 MCP tools to record trades across five memory layers (episodic, semantic, procedural, affective, trade records) and recall them with outcome-weighted scoring. Before trading, query recall_memories to see past setups. After, call remember_trade to log everything. You get confidence tracking, drawdown alerts, and SHA-256 audit trails for compliance. Works with any broker or market since it only stores decisions, never executes. Particularly useful if you're running pre-flight checklists or need MiFID II documentation for algorithmic trading decisions.
claude mcp add --transport stdio mnemox-ai-tradememory-protocol uvx tradememory-protocol