Paste your OpenAPI spec or JSON payload. Get dialect-correct SQL DDL, vector schemas, and dbt source models — across 8 databases, in under a second. Every time.
No sign-up · no credit card · 10 free generations
SQL DDL
Dialect-correct CREATE TABLE for every SQL warehouse.
CREATE TABLE payments ( id VARCHAR NOT NULL, amount FLOAT, currency VARCHAR(3), created_at TIMESTAMP_NTZ );
Vector Schema
Index config and metadata fields for vector databases.
# Pinecone — 1536 dims
index.upsert([{
"id": "pay_123",
"values": embedding,
"metadata": {
"amount": 99.9,
"currency": "GBP"
}
}])dbt Source
PROsource.yml with column tests + staging SQL, ready to commit.
- name: id
tests: [not_null, unique]
- name: currency
tests:
- accepted_values:
values: [USD,EUR,GBP]
- name: created_at
description: ""REST API or Python SDK — SQL DDL, vector schemas, and dbt source models, deterministic in every environment.
curl -X POST https://api.schemaforge.dev/api/v1/generate \
-H "Authorization: Bearer sf_live_..." \
-H "Content-Type: application/json" \
-d '{
"input_type": "openapi",
"content": "$(cat openapi.yaml)",
"output_type": "sql_ddl",
"sql_destination": "snowflake"
}'{
"output": "CREATE TABLE payments (\n id VARCHAR NOT NULL,\n ...",
"table_name": "payments",
"fields_detected": 6,
"warnings": [],
"generation_id": "d4f7b2c1-..."
}Same input → same output every run. No AI variance. Safe to pin in CI/CD.
pip install schemaforge-sdk
"output_type": "dbt_source" → source.yml + stg_*.sql
POST /generate · GET /history · GET /usage