Natural Language ORM · Node.js & TypeScript

Your database
speaks English
now.

Write a sentence. Get a real query.
Zero SQL. Zero boilerplate.
Compiles once — runs at full speed forever.

8+Databases
0msRuntime AI
1Compile step
Query speed
user.service.tsdb.config.ts
1import { MaskDatabases } from 'mask-databases';
2
3// ✨ Write in plain English
4const result = await MaskDatabases.prompt("fetch all users");
5console.log(result);
6
console.log(result)
[ { _id: "64f3a1…", name: "Alice", email: "alice@co.com" },
  { _id: "64f3a2…", name: "Bob", email: "bob@co.com" }, …2 more
]
Zero runtime AICompiled ahead of time
🗄️
8+ databasesMongoDB · PostgreSQL · MySQL…
🧠
Schema-awareKnows your exact table shapes
🗣️
Just EnglishNo query syntax to learn
How it works

Write once. Compile once. Query fast.

No runtime AI. No unpredictable output. Your queries are real, deterministic database code.

01

Define your models

Describe your schema in plain English. Mask turns it into real schemas for your DB engine.

MaskModels.define("User…")
02

Write prompts

Replace SQL or query builders with natural language. Say what you want — Mask handles the syntax.

MaskDatabases.prompt("fetch all users")
03

Run the compiler

One CLI command turns every prompt into production-ready, executable database code.

npx mask compile
Features

Everything your data layer needs

From schema definition to live queries — Mask handles the whole ORM layer.

🗣️

Stop Writing Queries. Just Say It.

Describe your intent in plain English. No SQL syntax, no aggregation pipelines — just say what you need and let the compiler handle the rest.

📖

Readable Like English. Runs Like Code.

Your queries read like documentation and execute like compiled code. Reviews, onboarding, and late-night debugging all get dramatically easier.

🗄️

Learn Any Database Instantly

MongoDB, PostgreSQL, MySQL, SQLite, Neo4j — same English interface for every engine. Switch databases without relearning a new query language.

🐛

Goodbye Query Debugging

Queries are compiled once and locked in. No runtime surprises, no AI unpredictability — just deterministic, production-safe database calls every time.

🧠

Your Database… Understood.

The compiler knows your exact schema — table names, field types, relationships. Every generated query is accurate to your data model, not a generic template.

👥

Built for Teams, Not Just You

Compiled output syncs across your whole team and CI pipelines. Everyone always runs the same queries — no drift between branches or deployments.

Supported databases

MongoDBPostgreSQLMySQLSQLiteMariaDBOracleNeo4jMongoose
Before vs After

Stop writing queries.
Start writing English.

Same result. One tenth of the code. Readable in a code review at midnight.

❌ Before Mask Databases Verbose
// MongoDB raw query

const users = await User
  .find({ status: 'active', role: 'admin' })
  .select('name email createdAt')
  .sort({ createdAt: -1 })
  .limit(50)
  .lean();
✅ With Mask Databases Clean
// Plain English

const users = await MaskDatabases.prompt(
  "get active admin users,
   name and email, newest first,
   limit 50"
);

// Compiled to the exact query above ↑

Ship the feature.
Not the query.

No raw queries, ever.

FAQ

Frequently asked questions

Everything you need to know before getting started.

Mask Databases supports MongoDB, Mongoose, MySQL, PostgreSQL, SQLite, MariaDB, Oracle, and Neo4j. More adapters are added regularly.

No. You write queries in plain English — for example "find all users with status active". The compiler converts your natural language into the correct database query automatically.

You define models and queries in plain English using MaskModels.define() and MaskDatabase.prompt(). You run the compiler once — it converts your prompts into real database code. At runtime there are no AI calls, so queries are fast, predictable, and production-safe.

Mask Databases currently supports Node.js and TypeScript via the mask-databases npm package. Install it with npm install mask-databases.

Plans start at $5/month. See the pricing page for all available options.