#1 Persistent memory for AI coding agents based on real-world benchmarks
TypeScript agentmemoryagentsaiclaudeclaudecodecodexcopilotcursorgenaiharnesshermesmemoryompopenclawopencodepi
{
"name": "agentmemory",
"version": "0.9.31",
"description": "Persistent memory for Cursor: auto-captures sessions, prompts, and tool use via hooks, recalls with hybrid BM25 + vector + graph search, and exposes 54 memory tools over MCP. Keyless by default, runs entirely on your machine.",
"author": {
"name": "Rohit Ghumare",
"email": "[email protected]"
},
"homepage": "https://agent-memory.dev",
"repository": "https://github.com/rohitg00/agentmemory",
"license": "Apache-2.0",
"keywords": ["memory", "mcp", "hooks", "recall", "knowledge-graph", "agent-memory"],
"logo": "assets/logo.svg",
"skills": "plugin/skills/",
"hooks": "plugin/cursor/hooks.json",
"mcpServers": "plugin/cursor/mcp.json",
"variables": {
"type": "object",
"properties": {
"AGENTMEMORY_URL": {
"type": "string",
"title": "agentmemory server URL",
"description": "REST base URL of the running agentmemory server",
"default": "http://localhost:3111"
},
"AGENTMEMORY_SECRET": {
"type": "string",
"title": "agentmemory secret",
"description": "Bearer token, required only when the server sets AGENTMEMORY_SECRET",
"default": ""
}
}
}
}