DB to API

Write a SELECT, an INSERT, or call a stored procedure. Apinizer publishes it as a REST endpoint with parameter binding, type coercion, audit, and the same security policies as the rest of your gateway.

API Creator · DB to API

Create powerful REST APIs instantly from your database without any coding.

Request a demo · Read the docs

Highlights

Databases — Oracle · PostgreSQL · MySQL · MSSQL · DB2 · Sybase · Hive · Impala · Trino · MongoDB
Operations — SELECT · INSERT · UPDATE · DELETE · PROCEDURE · BATCH
Pool — HikariCP · vault-encrypted credentials

Capabilities

01 · Bring your database. Apinizer brings the driver.

Ten databases ship with the platform — drivers, pool tuning, cursor unwrapping, and OUT-parameter handling are all built in. Add any other JDBC database by dropping its driver jar in.

Oracle, PostgreSQL, MySQL, SQL Server, IBM DB2, SAP Sybase
Apache Hive, Apache Impala, Trino — for warehouse and lake reads
MongoDB via the official driver (find-style queries)
JDBC-compatible? Drop the jar in, set credentials, you are done

Concepts: JDBC · MongoDB · HikariCP · Vault secrets

02 · Connect once. Reuse everywhere.

Pick a driver, paste the JDBC URL, store credentials in the platform vault. Apinizer test-pings the database, sizes the connection pool, and surfaces it to every endpoint that needs it.

JDBC URL or hostname / port / database fields
Credentials encrypted via Apinizer's secret vault — never in code
HikariCP pool with min / max / idle tuning per datasource
Optional read-replica routing for read endpoints

Concepts: test connection · pool tuning · read replica

03 · Write the query. Apinizer wires the route.

Type the SQL you would type in a client. Use named parameters like :customerId or :from. Map each parameter to a path segment, query string, header, or body field. Save — the endpoint is live.

Named parameters (:name) — never string-concatenated, always prepared
Per-parameter source: PATH · QUERY · HEADER · BODY
Type-aware binding: STRING, INT, LONG, DOUBLE, BOOLEAN, DATE, DATETIME, CLOB
JSON or XML response — toggle, never write a serializer

Concepts: SELECT · INSERT · UPDATE · DELETE · TRUNCATE

DB-to-API endpoints are first-class proxies — an AI agent can hit them through MCP just like any other API surface, with the same auth and audit.

04 · Multiple result sets and OUT parameters — in one JSON.

Procedures with REF_CURSOR results and OUT parameters return as a single, predictable JSON body. Apinizer handles the driver-specific cursor codes for Oracle, DB2, PostgreSQL, MySQL, SQL Server, and Sybase so your callers do not have to.

IN, OUT, INOUT direction per parameter
REF_CURSOR results unwrapped to JSON arrays
Multiple result sets walked via getMoreResults
Per-method query timeout (defaults to 30 seconds)

Concepts: REF_CURSOR · OUT params · Oracle · DB2 · Postgres · MSSQL

05 · One JSON array in. One executeBatch on the wire.

Bulk-loading orders, ledger entries, telemetry rows? Drop a JSON array on /endpoint/bulk and Apinizer turns it into a single executeBatch — with row count, error reporting, and the same query timeout you set on every other call.

BATCH_INSERT reads $.data[] from the request body
Returns row count and elapsed time in a structured response
executeBatch on the JDBC driver — not row-by-row inserts
Same parameter binding, type coercion, and audit as single-row calls

Concepts: BATCH_INSERT · executeBatch · $.data[]

06 · Strip nulls. Hide empty rows. Force big integers as strings.

Database results rarely match the JSON your callers want. Apinizer ships a response shaper so you don't write a thin service just to clean up nulls or coerce dates to ISO 8601. Toggle, save, ship.

Ignore NULL fields — drops keys whose value is null
Ignore empty rows — skips rows where every column is null
Numbers as strings — avoids JS precision loss for IDs > 2^53
Dates serialised as ISO 8601 in UTC — no caller-side parsing surprises

Concepts: JSON · XML · ISO 8601 · no precision loss

07 · A SQL endpoint is a first-class proxy.

DB-to-API endpoints are not second-class. They go through the same Worker, the same auth methods, the same throttling, the same cache, and the same Elasticsearch analytics as your hand-written services. Zero extra wiring.

OAuth 2.0 / JWT / API key / mTLS — all supported on DB endpoints
Throttling, quotas, IP allow-lists, header-based rate limits
Cache TTL or coordinated invalidation through Apinizer Cache
Every call audit-logged and indexed in Elasticsearch

Concepts: OAuth · throttle · cache · audit · Elasticsearch

Use cases

Expose database data without a microservice

Turn the table that backs a legacy system into a modern REST endpoint. No new service to run, no pod to schedule.

Stored procedure as endpoint with cursor results
Read replica routing for read-heavy paths
Per-endpoint quotas and cache TTL
Same audit trail as hand-written services

Hand the warehouse to product teams without copying it

Hive, Impala, Trino — front the warehouse with a tight, parameterised endpoint instead of giving every team query access.

Scope each query to a parameter the caller must supply
Throttle per consumer to protect the cluster
Cache repeat reads with a coordinated invalidation
Audit who pulled what, when, and why

Accept high-volume writes without a custom ingest service

Logistics, telemetry, ledger entries — accept JSON arrays and dispatch them as a single executeBatch.

BATCH_INSERT with row-count response
Per-method query timeout to protect the database
OAuth-scoped writers, IP allow-lists, mTLS
Drop in dead-letter handling via response policies

What ships in the box

Databases & operations

Oracle, PostgreSQL, MySQL, SQL Server, IBM DB2, SAP Sybase, Apache Hive, Apache Impala, Trino, MongoDB
SELECT, INSERT, UPDATE, DELETE, PROCEDURE, TRUNCATE, BATCH_INSERT (JDBC drivers)
MongoDB find-style queries with placeholders
Drop in any other JDBC driver jar to add a database

Parameters & shaping

Named parameters (:name) bound from PATH / QUERY / HEADER / BODY
Per-parameter direction (IN / OUT / INOUT) and type
REF_CURSOR results and OUT parameters in one JSON response
Per-endpoint query timeout (default 30 s) — protects the database
Response shaping: ignore null, ignore empty, numbers as strings, ISO 8601 dates

Pipeline integration

Same auth surface as gateway proxies (OAuth 2.0, JWT, API key, mTLS, Basic)
Throttling, quotas, IP allow-lists, scope-based access
Cache TTL with coordinated invalidation across replicas
Audit log + Elasticsearch analytics + Prometheus metrics
APIops manifests so DB endpoints ship through CI/CD like any other proxy

Resources

DB-to-API docs — Connection setup, parameter binding, stored procedures, batch operations.
Cursor & OUT parameter guide — Driver-specific notes for Oracle, DB2, PostgreSQL, MySQL, SQL Server, Sybase.
Mock API — Test the call shape before the real query is wired.
Architecture — How DB-to-API endpoints sit on Worker and pass the policy stack.

Next step

Stop writing translation services

Some endpoints are just a SELECT. Stop building services for them.

See DB-to-API on a real database in a 30-minute walkthrough — including stored procedures, batch loads, and the policy pipeline.

Book a Demo · Read the docs