@booboo-brain/serve

@booboo-brain/serve

Query a Booboo snapshot over REST and MCP. One in-memory index (BoobooIndex) powers both: search, neighbours, shortest-path, layer stats. Read-only.

REST

booboo serve --snapshot my.booboo.json --port 8787

Route	Returns
GET /graph	meta + counts
GET /stats	{ nodes, links, byLayer }
GET /search?q=&limit=	ranked nodes (exact > prefix > substring)
GET /nodes?layer=&cluster=&type=&q=&limit=&offset=	{ total, nodes }
GET /nodes/:id	one node (404 if missing)
GET /neighbors/:id?depth=&limit=	{ center, nodes, links }
GET /path/:from/:to	{ path } (chain of nodes, or null)

CORS is open, so a browser viewer can fetch a live graph directly.

MCP

booboo mcp --snapshot my.booboo.json

Speaks MCP over stdio. Point any client at it (Claude Desktop / Claude Code):

{ "mcpServers": { "booboo": { "command": "booboo", "args": ["mcp", "--snapshot", "/abs/path/my.booboo.json"] } } }

Tools: booboo_stats · booboo_search · booboo_node · booboo_neighbors · booboo_path.

API

import { BoobooIndex, loadSnapshot, createRestServer, runMcp } from "@booboo-brain/serve";

const ix = new BoobooIndex(loadSnapshot("my.booboo.json"));
ix.search("invoice");           // ranked nodes
ix.neighbors("dionisos", 2);    // 2-hop neighbourhood
ix.path("a", "b");              // shortest path or null
createRestServer(ix).listen(8787);

BoobooIndex is pure (no Node/network deps), so it also runs in the browser.