MCP server that caches expensive LLM results — HTTP-style freshness plus semantic recall of related answers
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add echocache -- npx -y echocache~/.codex/config.toml
[mcp_servers.echocache]
command = "npx"
args = ["-y", "echocache"]
# codex mcp add echocache -- npx -y echocache~/.cursor/mcp.json
{
"mcpServers": {
"echocache": {
"command": "npx",
"args": [
"-y",
"echocache"
]
}
}
}.vscode/mcp.json
{
"servers": {
"echocache": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"echocache"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"echocache": {
"command": "npx",
"args": [
"-y",
"echocache"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"echocache": {
"command": "npx",
"args": [
"-y",
"echocache"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"echocache": {
"command": "npx",
"args": [
"-y",
"echocache"
]
}
}
}