Self-hosted MCP server + dashboard giving Claude Code, Cursor & Hermes agents shared long-term memory over a git-backed markdown vault. Agent-native, Obsidian-compatible second brain — markdown RAG, no vector DB.
TypeScript agent-memoryai-agentsanthropicclaude-codecursorgitknowledge-baseknowledge-graphllmmarkdownmcpmcp-servermodel-context-protocolnote-takingobsidianpkmragsecond-brainself-hostedzettelkasten
import { AUTH_DISABLED, AUTH_SECRET, MCP_TOKEN } from "@/lib/config";
import { appName, curatorMode, harnessEnabled } from "@/lib/settings";
import { hasAnyToken } from "@/lib/tokens";
export const dynamic = "force-dynamic";
export function GET() {
return Response.json({
appName: appName(),
/** "off" | "chat" | "full" — whether Engram itself runs a model, and whether it may write. */
curator: curatorMode(),
/** Kept for the Connect page: brain_capture is exposed only in "full". */
harness: harnessEnabled(),
mcpAuthRequired: MCP_TOKEN !== "" || hasAnyToken(),
dashboardAuthRequired: !AUTH_DISABLED && AUTH_SECRET !== "",
});
}