Air Pipe speaks to relational and document databases natively. Write a query, add assertions and auth, and you have a production endpoint — no ORM, no Mongoose, no backend boilerplate.
Relational and document, first-class
PostgreSQL, MySQL/MariaDB, SQL Server and SQLite for relational; MongoDB for documents, including aggregation pipelines and vector search. Your data stays where it is — Air Pipe connects to it.
SQL CRUD and parameterised queries as config
Mongo document operations and aggregation
pgvector and Atlas Vector Search for RAG
No ORM, no glue code
There's no model layer to maintain and keep in sync. The query is the source of truth; assertions validate input; transforms shape output. That's the whole backend.
Secured and observable out of the box
Every database-backed endpoint gets auth, validation, tracing, metrics and OpenAPI docs automatically — the same as any other Air Pipe interface.
Frequently asked questions
Which databases does Air Pipe support? PostgreSQL, MySQL/MariaDB, SQL Server and SQLite for relational work, and MongoDB for documents — including aggregation pipelines and vector search. They stay where they are; Air Pipe connects to them.
Do I need an ORM? No. You write the query you want in config, with parameters and assertions around it. There's no model layer to define and no migration framework to adopt.
Can Air Pipe connect to a database on my own network? Yes. Self-host the binary alongside the database and the connection never leaves your network — the config is identical to the managed one.
Is it safe against SQL injection? Queries take bound parameters rather than string-concatenated input, and you can assert on any input before it reaches the query.