@monlite/postgres

The Postgres engine for monlite. The same @monlite/core API — documents, queries, the full CRUD surface — on a networked Postgres (documents stored as JSONB) instead of a local SQLite file. Swap the engine, not your code.

npm install @monlite/core @monlite/postgres pg

import { createDb } from "@monlite/postgres";

const db = createDb("postgres://user@host/db");

await db.collection("users").create({ data: { name: "Ada", age: 30, tags: ["admin"] } });
const adults = await db.collection("users").findMany({
  where: { age: { gte: 18 }, tags: { has: "admin" } },
  orderBy: { age: "desc" },
});

Or pass the engine to core's createDb:

import { createDb } from "@monlite/core";
import { postgres } from "@monlite/postgres";

const db = createDb("pg", { driver: postgres("postgres://…", { pool: { max: 10 } }) });

Same API, different engine

Documents become jsonb; the query builder emits the Postgres dialect (->/->>/@>/~, jsonb_array_elements, to_jsonb); transactions run on a pooled client with SAVEPOINTs. The db.collection(...) API is identical to SQLite-backed monlite — develop on a .db file, deploy to Postgres, don't rewrite a line.

The whole data surface works:

• CRUD — create, createMany, findMany, findFirst, findById, count, exists, update, updateMany, upsert, delete, deleteMany, findOneAndUpdate, bulkWrite, purgeExpired.

• Aggregation — aggregate, groupBy (with having + orderBy), distinct.

• Realtime — watch() via Postgres LISTEN/NOTIFY (truly cross-process — a write from any connection reaches every watcher).

• Full-text search — @monlite/fts on a native generated tsvector column + GIN index.

• Vector search — @monlite/vector on a native generated vector column + HNSW index (pgvector).

• Job queue — @monlite/queue's createPgQueue(db), claiming with FOR UPDATE SKIP LOCKED.

Not yet: only explain() (Postgres' EXPLAIN output is engine-specific — throws a clear error).

Notes

• Requires PostgreSQL 14+ (vector search needs pgvector; the monlite/postgres image bundles it). Placeholders ? are rewritten to $1,$2,….
• Each transaction runs on its own pooled connection — concurrent transactions parallelize (nested ones become SAVEPOINTs) and retry automatically on serialization failure / deadlock.
• @monlite/core stays the minimal, zero-dependency local engine — this package is purely additive, opt-in.

License

MIT