Core concepts

Hyperterse is built around three simple concepts: Adapters, Queries, and Inputs. Understanding these will help you build powerful data APIs quickly.

Key concepts

Adapters Database connections. Define how Hyperterse connects to PostgreSQL, MySQL, or Redis.

Queries SQL statements that become API endpoints. Define what data to expose.

Inputs Typed parameters for your queries. Define what data clients can pass in.

How they work together

1. Adapters establish database connections
2. Queries define SQL statements that use those adapters
3. Inputs parameterize queries with validated, typed data
4. Hyperterse generates REST endpoints, MCP tools, and documentation automatically

Example

name: my-api

# 1. adapter: connect to the database
adapters:
  main_db:
    connector: postgres
    connection_string: 'postgresql://user:pass@localhost:5432/app'

# 2. query: define what to expose
queries:
  get-user:
    use: main_db # Which adapter to use
    description: 'Get user by ID'
    statement: |
      SELECT id, name, email
      FROM users
      WHERE id = {{ inputs.userId }}  # 3. Input: Parameterize
    inputs:
      userId:
        type: int
        description: 'User ID'

This single configuration generates:

• POST /query/get-user — REST endpoint
• MCP tool get-user — For AI assistants
• OpenAPI documentation at /docs
• LLM-friendly documentation at /llms.txt

What Hyperterse is not

Understanding what Hyperterse doesn’t do helps set expectations:

Hyperterse Is	Hyperterse Is Not
A query gateway	An ORM
Configuration-driven	A code generator
Stateless runtime	A database migration tool
Read-optimized	A full CRUD framework

You write SQL directly—Hyperterse doesn’t abstract it away. This gives you full control over query performance and behavior.