API as Architecture
Embrace AI with every API.
One contract. Your APIs, services, databases — a design language that embraces both engineers and AI agents.
$ iex (irm 'https://raw.githubusercontent.com/apihug/apihug.github.io/main/helper/apihug-install.ps1')syntax = "proto3";service OrderService { rpc Place () returns () { }; }}public class OrderController { @PostMapping("/order/place") public place() { // Generated by ApiHug }}The Blueprint
The semantic execution infrastructure for AI-native enterprises.
From AI agents at the top to enterprise systems at the bottom — ApiHug is the operating language that connects them all, flanked by security, governance, observability, and collaboration.
Proto Semantic
Protobuf is AI-native by design.
ApiHug extends protobuf with a rich semantic layer — routing, validation, persistence, mocking — all declared in your .proto files. Typed fields, pagination, repeated inputs — structured semantics that LLMs understand without hallucinating.
HTTP routes, methods, pagination — declared on your RPCs. Generates OpenAPI specs, Spring controllers, and client SDKs automatically.
service OrderService { // Service-level config option (hope.swagger.svc) = { path: "/order"; description: "Order Service"; } rpc Place (PlaceOrderRequest) returns (OrderPlacedResponse) { option (hope.swagger.operation) = { post: "/place"; description: "place a new order"; }; } rpc Query (QueryOrderRequest) returns (OrderView) { option (hope.swagger.operation) = { get: "/query"; pageable: true; }; }}message Movie { string name = 1 [(hope.swagger.field) = { description: "name of the movie"; example: "The Lord of the Rings" empty: FALSE; mock: { nature: MOVIE } max_length: { value: 64 } }]; MovieLevel level = 2 [(hope.swagger.field) = { description: "level of the movie"; example: "PG_13" empty: FALSE; }]; uint32 year = 3 [(hope.swagger.field) = { description: "publish year"; example: "2022" }];}enum MovieLevel { PG_13 = 0 [(hope.constant.field) = { code: 1, message: "PG-13" }]; R = 1 [(hope.constant.field) = { code: 2, message: "R" }]; NC_17 = 3 [(hope.constant.field) = { code: 4, message: "NC-17" }];}// 40+ built-in nature types for mock dataenum Nature { ANIMAL = 0; // random animal names AVATAR = 2; // avatar image URLs BOOK = 4; // book titles COLOR = 8; // color names COUNTRY = 10; // country names CURRENCY = 11; // currency names DATE_AND_TIME = 12; // random dates DOG = 13; // dog breed names FOOD = 14; // food names MOVIE = 20; // movie titles NAME = 22; // person names EMAIL = 24; // email addresses PHONE = 26; // phone numbers URL = 36; // random URLs}Entity Design
From proto to database, automatically.
Define your domain model in protobuf. ApiHug generates entities, migrations, and CRUD — schema, code, and API always in sync.
message Movie { string name = 1 [(hope.persistence.column) = { name: "NAME", description: "Name of the movie", type: VARCHAR, length: { value: 64 } }]; string description = 2 [(hope.persistence.column) = { name: "DESCRIPTION", type: VARCHAR, length: { value: 255 } }]; MovieLevel level = 3 [(hope.persistence.column) = { enum_type: STRING, type: VARCHAR, length: { value: 16 } }]; option (hope.persistence.table) = { name: "MOVIE", wires: [IDENTIFIABLE, AUDITABLE] };}<!-- Auto-generated by ApiHug --><createTable remarks="Movie" tableName="MOVIE"> <column name="ID" type="BIGINT" autoIncrement="true"> <constraints primaryKey="true" unique="true" /> </column> <column name="NAME" type="VARCHAR(64)"> <constraints nullable="false" /> </column> <column name="DESCRIPTION" type="VARCHAR(255)" /> <column name="LEVEL" type="VARCHAR(16)" /> <column name="CREATED" type="TIMESTAMP" /></createTable>@Entity@Table(name = "MOVIE")@Auditablepublic class Movie implements Identifiable<Long> { @Id @GeneratedValue private Long id; @Column(name = "NAME", length = 64, nullable = false) private String name; @Column(name = "DESCRIPTION", length = 255) private String description; @Enumerated(EnumType.STRING) @Column(name = "LEVEL", length = 16) private MovieLevel level;}Enterprise Factory
One definition. Everything generated.
No more hand-wiring layers. One proto contract generates services, database, SDKs, testing, and AI tooling — the same definition powers both human engineers and machine agents.
Protobuf Contract
Describe your API in Protocol Buffers — structured, typed, machine-readable.
Spring Boot Services
Controllers, DTOs, and service stubs — production-ready, zero boilerplate.
Database Schema
Tables, indexes, and migrations — generated from your entity annotations.
MCP & Agent Tools
MCP servers and tool schemas so AI agents can invoke your APIs safely.
Agent Native
Embrace the age of AI agents.
In the AI era, your API consumers are agents — not just developers. ApiHug hugs both: giving humans clean contracts and LLMs the structured semantics they need to understand — not guess at — your enterprise.
MCP Server Generation
Auto-generate MCP servers from your API contracts. AI agents discover and invoke services without custom integration.
Tool Schema & Catalog
Every endpoint becomes a typed tool definition with input, output, and pagination — exactly what LLMs need.
Agent Permission Model
RBAC and audit trails keep AI agents operating within safe enterprise boundaries.
Semantic API Contracts
Protobuf eliminates ambiguity. LLMs reason best against structured, typed definitions with clear semantics.
SDK & Language Bindings
Client SDKs for any language from one contract — Java, TypeScript, Python, Go.
Enterprise Governance
Observability, tracing, and compliance built into every generated artifact.
Enterprise Intent Compiler
From business intent to deployed systems. ApiHug hugs both humans and AI agents into a shared, unambiguous contract — the foundation for the entire enterprise lifecycle.