Skip to content
Axel

Exactly how SQL should be written

Axel
Announcing Axel — Ahead-of-Time PostgreSQL Compiler

SQL generation for schemas + queries

Write schemas in ASL, queries in AQL. Axel compiles both to clean PostgreSQL SQL — migrations and parameterized query strings. No ORM, no driver, no runtime magic.

Get Started
DocumentationSchema (ASL)Queries (AQL)Reference
v1.0

Quick look

Copy code

Define models in ASL, query shapes in AQL, and get optimal single-scan PostgreSQL queries.

use extension 'pgcrypto';

type User {
  required id: uuid { default := gen_uuid(); constraint pk; };
  required email: str { constraint exclusive; };
  name: str;
}

type Post {
  required title: str;
  required link author: User;
  multi link likes: User;
}
Built for teams creating modern, high-performance PostgreSQL applications with zero runtime overhead.
Schema Language (ASL)

Declarative Schemas, Zero Decorator Gymnastics

Define your models, single/multi links, indexes, and custom constraints cleanly in .asl files. Axel AST-diffs your schema and generates clean, deterministic PostgreSQL migration SQL automatically.

Explore Schema Language (ASL)
schema.asl
ASL Model
type Article extending Base {
  required title: str;
  slug: str { constraint exclusive; };
  required link author: User;
  multi link tags: Tag;
}
Query Compiler (AQL)

Nested Shapes in a Single Database Scan

Select deeply nested relational graphs in one intuitive query. Axel compiles nested shapes into a single PostgreSQL round-trip — zero N+1 queries. Choose between two loading strategies:

  • query (default) — correlated subqueries with json_agg and row_to_json in the SELECT projection.
  • join — LEFT JOIN LATERAL in the FROM clause, ideal when the planner benefits from seeing all joins together.

Set globally in axel.yaml or override per-query with @rel_load_strategy.

Explore Nested Shapes
get_article.aql
AQL Query
select Article {
  id,
  title,
  author: { id, email },
  tags: { id, name }
}
filter .slug = $slug;

-- compiles to json_agg + row_to_json subqueries
Ahead-of-Time Architecture

Ahead-of-Time Compiler, Zero Runtime Overhead

Axel is a build-time compiler producing pure parameterized SQL strings and typed structs/interfaces. There is no query engine sidecar, no connection pooling layer, and zero driver wrapping.

Explore Code Generation
Runtime Architecture
Zero Footprint
✓ Axel (AOT):Compiles at build-time to plain SQL strings. Zero runtime CPU overhead, zero connection wrapper, zero cold-starts.
✗ Heavy Runtime Engines:Bundle Rust binaries (Prisma 6) or WASM compilers (Prisma 7) adding memory overhead and locking execution to Node.js.
PostgreSQL Superpowers

Extensions, Row-Level Security, Conflicts & Triggers

Axel is purpose-built for PostgreSQL. Declare database extensions (pgvector, pg_trgm), fine-grained Row-Level Security (policy), triggers, and unless conflict upsert semantics directly in your schema and queries.

Explore Policies & RLS Guide
policies.asl
PostgreSQL Native
policy tenant_isolation on Organization
  for all using (.id = global.current_org_id);

Bring your database
into the AOT age

Get started in seconds with Axel's CLI, schema modeling, and zero-runtime query compiler.

Get Started