prorm documentation
Choose your style
produce the same models, so everything below applies whichever you choose.
require and go. No build step, no types.ts-prorm-orm is a TypeScript ORM covering 77 SQL dialects, 6 graph querycompilers registered under 14 dialect names, and 105 additional data stores
(how each number is counted is in Architecture).
These pages are the developer-facing documentation: what each subsystem does,
how to use it, and the SQL it produces.
npm install ts-prorm-orm better-sqlite3pnpm add ts-prorm-orm better-sqlite3yarn add ts-prorm-orm better-sqlite3bun add ts-prorm-orm better-sqlite3getting started, or run
prorm packagesto pick the ones your project needs.
import { Prorm, DataTypes, Op } from 'ts-prorm-orm';
const prorm = new Prorm({ dialect: 'sqlite', storage: ':memory:' });
await prorm.connect();
const User = prorm.define('User', {
id: { type: DataTypes.INTEGER(), primaryKey: true, autoIncrement: true },
name: { type: DataTypes.STRING(100), allowNull: false },
email: { type: DataTypes.STRING(255), allowNull: false, unique: true },
age: { type: DataTypes.INTEGER() },
}, { tableName: 'users', timestamps: true });
await prorm.sync();
await User.create({ name: 'Ada', email: 'ada@example.com', age: 36 });
const adults = await User.findAll({ where: { age: { [Op.gte]: 18 } } });reach for:
Start here
define(), decorators, table options, timestamps, soft deletes, sync().DataTypes.* and its SQL on each dialect.typeMapping: 'legacy' vs 'modern', and which 4.8 columns need a manual migration.where, order, pagination, aggregates, writes — with the SQL.Op.* reference.pipe() and Q — Composing FindOptions step by step, and how the steps merge.hasOne, hasMany, belongsTo, belongsToMany.include, nested loads, filtering includes, N+1.Working with data
| Guide | What it covers |
|---|---|
| Bulk operations | bulkCreate, bulkUpdate, increment, batching |
| Transactions | Callback and manual modes, isolation levels, savepoints |
| Raw queries | prorm.query(), replacements, query types |
| Streaming | Cursors, paginated fallback, transform pipelines |
| Scopes | Reusable, composable query fragments |
| Hooks | The lifecycle events and when each fires |
| Validation | Field and model validators |
| Virtual fields | Computed attributes with no column |
| Error handling | The error hierarchy and how to branch on it |
Schema
| Guide | What it covers |
|---|---|
| Indexes & constraints | Declarative and runtime index/constraint management |
| Schema objects | Views, materialized views, sequences, triggers, procedures, RLS, partitions |
| Migrations | Migration files, the Migrator, seeders |
| QueryInterface | The imperative schema API migrations use |
| Schema diffing | Diff models against the live database, generate migrations |
| Prisma import | Convert a schema.prisma into models and a migration |
| Models from a database | prorm models pull: generate models from any live SQL database, every dialect |
Connections & scale
| Guide | What it covers |
|---|---|
| Connection pooling | Pool sizing, lifecycle, events |
| Multiple databases | ConnectionManager, cross-database queries |
| Read replicas | Routing strategies, lag detection, health |
| Caching | L1/L2 cache, Redis cluster backend, invalidation |
| Query optimization | EXPLAIN plans, hints, slow-query logging, batching |
| Logging | Query logging, the Logger, query events |
Extending the model layer
| Guide | What it covers |
|---|---|
| Decorators | The class-based API and the full decorator families |
| TypeScript types | InferAttributes, CreationOptional, ModelStatic |
| External fields | Column contents stored in S3, Redis, GCS, … |
| Enums and constants | The named vocabularies — constraint types, referential actions, isolation levels, index methods |
| SQL constants | Isolation levels, PRAGMAs, sql_mode, search path |
| Extensions | Declare extensions so sync() installs them; prorm extensions list/import/export; managed-service allow-lists; the 124-entry catalogue |
| Local databases | prorm db up --dialect postgres --ext pgvector,postgis: a pinned image, generated credentials, Docker or Podman |
Operations, security & compliance
| Guide | What it covers |
|---|---|
| Audit logging | Row-level change history |
| Audit levels & escalation | Levels on audit entries (RFC 5424 + outage, custom ladders), escalation policies as JSON, 30 presets, notifiers, aggregation, retention by level |
| StoreKey — passwordless prod-orm | Reach a prod-orm database with a .klock instead of a connection; described queries, TLS and Kerberos identity |
| Query firewall | Allow/block described as preset lists — allow('CRUD'), block(['DESTRUCTIVE', 'OWASP']) — instead of raw regular expressions |
| User management | Database users, roles, GRANT / REVOKE |
| Compliance | Jurisdictions, frameworks, encryption, masking, retention, erasure, consent, residency, the sheet |
| Foreign data wrappers | Querying other servers from PostgreSQL |
| Runbooks | Backup/restore, failover, rollback, on-call, DR |
| Known issues | Security advisories (PRORM-SA ids): risk level, likelihood, status, and the version and date each was resolved; how to report a vulnerability |
| Versions and their status | Every published version: current, deprecated or unsupported, the support policy, and how to upgrade |
| Security | Security overview, with detailed advisories for each fixed critical, high and medium issue: cause, impact, fix commit, how to verify and upgrade |
Databases
| Reference | What it covers |
|---|---|
| Database types | The full taxonomy — 22 types, 129 engines |
| SQL dialects | One page per dialect |
| Store adapters | 105 object, KV, document, search, queue, log and vector stores |
| Graph databases | Neo4j, Neptune, Gremlin, Dgraph through the model API |
| SPARQL / RDF triplestores | Stardog, Fuseki, Blazegraph, GraphDB, Virtuoso, AllegroGraph through the same graph dialect — the triple-mapping convention, and what SPARQL cannot express |
| ArangoDB (AQL) | Collections as labels, edge collections as edges, AQL compiled from ORM specs |
| Couchbase (N1QL) | Keyspace quoting, USE KEYS, the missing-index trap, and N1QL compiled from findAll-shaped specs |
| Loki (LogQL) | Stream selectors, pipelines, parsers, unwraps and metric aggregations compiled from LogQL-shaped specs — and the two string-quoting forms that keep them safe |
| InfluxDB (Flux) | A pipeline spec rather than a findAll, every stage compiled, and why escaping is the only safety net |
| Prometheus / VictoriaMetrics (PromQL) | Selectors, matchers, aggregations, the rate family and subqueries as a PromQL-shaped spec — plus what is escaped, what is refused, and where MetricsQL diverges |
| Log search (SPL, Lucene, Papertrail, Sumo) | Four log languages compiled from structured specs — how much they genuinely share, the per-language escaping and wildcard rules, and what each one refuses |
| SQLite advanced | FTS5, JSON1, generated columns, PRAGMA tuning |
Tooling
| Reference | What it covers |
|---|---|
| Seeders | Reference data and dev fixtures, with environments kept apart |
| Passwords | Hashing a password column with @Secure, and verifying one at sign-in |
| MCP server | The documentation as tools an AI agent can call |
| CLI | prorm — scaffolding, model generation, migrations, live-database inspection and maintenance, cloud commands |
| Terraform | The gitlab.com/prod-orm/prorm/restapi module: a database, its stores, pools, keys, users, network access, rules, roles and policies — declared, in a team project |
| In the browser | The CDN build: SQLite in the page, web storage, cookies, PouchDB/RxDB |
| Diagrams | SVG ER, class, model, migration and 11 other diagram generators |
| Docker | Running each engine locally in Docker |
Architecture references
connection, and what wraps that path. Code-level documentation lives beside
the source it describes:
dialects — the Dialect contract and per-engine implementationsmodels — the model layerquery-builders — where/order/include → SQL, and the typed SQL function libraryquery-optimizers — optimizer internalsmigrations — migration runner internalscompliance — per-file compliance status and known issuesdecorators — every decorator, option by optionextensions — the extension catalogue's shapeCONNECTIONS — pooling, connection/replica management, transactions, hooksUTILITIES — errors, validators, logging, audit, caching, diagrams, streamscli — how the CLI is put togetherA note on the SQL in these pages
sql blocks are the statements the ORM actually emits, captured from therunning code rather than written from memory, and are shown for SQLite unless a
page says otherwise — or offers a PostgreSQL / MySQL / SQLite / SQL Server tab,
which switch together across a page. Other dialects differ in quoting (
"col" / `col` /[col]) and placeholders (? / $1 / :1), and where a dialect doessomething materially different the page says so.
than documenting the intent. Those notes are marked as warnings in the
relevant guide, with the SQL that shows the behaviour and what to do instead.
Modified at 2026-10-04 00:40:16