Model Context Protocol server for Engrava — expose an AI agent memory database to any MCP client (Claude Desktop, Cursor, VS Code…) over stdio.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add engrava --env ENGRAVA_DB_PATH='<ENGRAVA_DB_PATH>' -- uvx engrava-mcp~/.codex/config.toml
[mcp_servers.engrava]
command = "uvx"
args = ["engrava-mcp"]
env = { ENGRAVA_DB_PATH = "<ENGRAVA_DB_PATH>" }
# codex mcp add engrava --env ENGRAVA_DB_PATH='<ENGRAVA_DB_PATH>' -- uvx engrava-mcp~/.cursor/mcp.json
{
"mcpServers": {
"engrava": {
"command": "uvx",
"args": [
"engrava-mcp"
],
"env": {
"ENGRAVA_DB_PATH": "<ENGRAVA_DB_PATH>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"engrava": {
"type": "stdio",
"command": "uvx",
"args": [
"engrava-mcp"
],
"env": {
"ENGRAVA_DB_PATH": "<ENGRAVA_DB_PATH>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"engrava": {
"command": "uvx",
"args": [
"engrava-mcp"
],
"env": {
"ENGRAVA_DB_PATH": "<ENGRAVA_DB_PATH>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"engrava": {
"command": "uvx",
"args": [
"engrava-mcp"
],
"env": {
"ENGRAVA_DB_PATH": "<ENGRAVA_DB_PATH>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"engrava": {
"command": "uvx",
"args": [
"engrava-mcp"
],
"env": {
"ENGRAVA_DB_PATH": "<ENGRAVA_DB_PATH>"
}
}
}
}Environment variables
ENGRAVA_DB_PATH
Replace the placeholders with your own credentials and keep them out of version control.