A local-first, token-efficient MCP memory server for AI coding clients
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add memory-lite -- npx -y memory-mcp-lite~/.codex/config.toml
[mcp_servers.memory-lite]
command = "npx"
args = ["-y", "memory-mcp-lite"]
# codex mcp add memory-lite -- npx -y memory-mcp-lite~/.cursor/mcp.json
{
"mcpServers": {
"memory-lite": {
"command": "npx",
"args": [
"-y",
"memory-mcp-lite"
]
}
}
}.vscode/mcp.json
{
"servers": {
"memory-lite": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"memory-mcp-lite"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"memory-lite": {
"command": "npx",
"args": [
"-y",
"memory-mcp-lite"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"memory-lite": {
"command": "npx",
"args": [
"-y",
"memory-mcp-lite"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"memory-lite": {
"command": "npx",
"args": [
"-y",
"memory-mcp-lite"
]
}
}
}