Connect Lore to Cursor, VS Code, Claude Code, Claude Desktop, Goose, Gemini CLI, Codex CLI, OpenClaw, and other MCP clients.
Lore exposes an MCP server that gives any compatible AI tool access to your knowledge base.
Before adding the MCP server, you must run the setup wizard to configure API keys and create your account:
npx @getlore/cli setup
This saves your API keys, auth session, and data directory to ~/.config/lore/. The MCP server reads this config automatically, so no env vars are needed in any of the configurations below.
Requires Node.js 18+.
One-command install (recommended):
# Add the Lore marketplace
/plugin marketplace add getlore-ai/claude-code-marketplace
# Install the plugin
/plugin install lore@getlore
This installs Lore's MCP server and a SKILL.md that teaches Claude when and how to use your knowledge base.
Manual config — add to .mcp.json in your project root, or ~/.claude.json for global access:
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"]
}
}
}
Open Settings > Developer > Edit Config and add:
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"]
}
}
}
One-command install:
gemini extensions install https://github.com/getlore-ai/lore --path plugins/gemini
This installs Lore's MCP server and a GEMINI.md context file with tool usage instructions.
Manual config — add to your Gemini settings:
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"]
}
}
}
Add the MCP server:
codex mcp add lore --command npx --args "-y,@getlore/cli,mcp"
Install the skill (optional — adds agent instructions):
lore skills install codex
Install from ClawHub:
clawdhub install lore
Or install via the Lore CLI:
lore skills install openclaw
Use the Install in Cursor badge above, or add manually to .cursor/mcp.json:
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"]
}
}
}
Use the Install in Goose badge above, or add via the CLI:
goose mcp add lore --command npx --args '-y @getlore/cli mcp'
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"]
}
}
}
All configurations above use npx, which downloads and runs the latest version automatically — no global install needed. If you prefer a global install:
npm install -g @getlore/cli
Then replace "command": "npx", "args": ["-y", "@getlore/cli", "mcp"] with "command": "lore", "args": ["mcp"] in any config above.
After configuring, verify the MCP server is working:
lore mcp
You should see the MCP server start and list 9 available tools. In your AI client, try asking:
Search my knowledge base for "getting started"
If the tool responds, you're all set.
You can override values from lore setup by adding an env block to any MCP config. Environment variables take precedence over ~/.config/lore/config.json. This is useful for running multiple accounts or custom Supabase instances.
{
"mcpServers": {
"lore": {
"command": "npx",
"args": ["-y", "@getlore/cli", "mcp"],
"env": {
"OPENAI_API_KEY": "different-key",
"LORE_DATA_DIR": "~/alternate-lore-data"
}
}
}
}
Available environment variables: OPENAI_API_KEY, ANTHROPIC_API_KEY, LORE_DATA_DIR, SUPABASE_URL, SUPABASE_PUBLISHABLE_KEY.