sekejap

Embedded graph-first, multi-model database for Node.js — SQL + graph + vector + spatial in one native library. No server, no external process; the whole engine runs in-process on a Rust core.

• PostgreSQL-like SQL — CREATE TABLE, INSERT, SELECT … WHERE, parameters ($1).
• Graph — link records and traverse relationships.
• Prepared statements — compile once, run many with varying parameters.
• No native build — the prebuilt binaries for every platform ship inside the package. No Rust toolchain, no node-gyp, no compile step.

Install

npm install sekejap

Works on macOS (arm64/x64), Linux (x64/arm64), and Windows (x64), Node ≥ 16.

Quick start

const { Db, version } = require('sekejap');

const db = Db.open('/tmp/notes');           // open (or create) a database directory

// Define a table and insert rows with SQL.
db.execute('CREATE TABLE note (_key TEXT PRIMARY KEY, title TEXT, pinned INTEGER)');
db.execute("INSERT INTO note (_key, title, pinned) VALUES ('n1', 'Buy milk', 1)");

// Insert a record directly (no SQL) as a JSON payload.
db.put('note/n2', JSON.stringify({ _collection: 'note', _key: 'n2', title: 'Call Sam', pinned: 0 }));

// Query. Results come back as a JSON string — JSON.parse it.
const notes = JSON.parse(db.query('SELECT * FROM note'));
console.log(notes);
// [ { _key: 'n1', title: 'Buy milk', pinned: 1, ... }, { _key: 'n2', ... } ]

// Parameterized query ($1, $2, …) — injection-safe. Params are a JSON array string.
const pinned = JSON.parse(db.queryParams('SELECT title FROM note WHERE pinned = $1', JSON.stringify([1])));

// Prepared statement — compile once, run many.
const byPin = db.prepare('SELECT _key FROM note WHERE pinned = $1');
for (const p of [0, 1]) {
  console.log(JSON.parse(db.queryPrepared(byPin, JSON.stringify([p]))));
}

// Graph: link two records.
db.link('note/n1', 'note/n2', 'related');
console.log('nodes:', db.nodeCount(), 'edges:', db.edgeCount());

// Flush + compact to disk after heavy writes.
db.compact();

API

| Method | Purpose | |---|---| | Db.open(path) | Open (or create) a database at a directory path | | db.execute(sql) | Run DDL/DML (CREATE/INSERT/UPDATE/DELETE) → rows affected | | db.query(sql) | Run a SELECT/graph query → JSON string | | db.queryParams(sql, paramsJson) | Parameterized query; paramsJson is a JSON array string | | db.prepare(sql) / db.queryPrepared(stmt, paramsJson) | Prepared statements | | db.put(slug, payloadJson) | Insert/replace one record by slug with a JSON payload | | db.link(from, to, edgeType) | Create a graph edge | | db.nodeCount() / db.edgeCount() | Counts | | db.compact() | Flush WAL, rewrite payloads, reclaim RAM | | version() | Library version |

query / queryParams / queryPrepared return a JSON string — decode with JSON.parse(...). It's an array of row objects.

TypeScript

Types ship with the package (index.d.ts) — no @types needed.

import { Db } from 'sekejap';
const db = Db.open('/tmp/notes');

License

Dual-licensed under MIT OR Apache-2.0.