Live demo · no signup

Describe the system once.DIS compiles the rest.

This runs the real compiler through the production gateway. Edit the spec or load a sample, hit compile, and inspect the generated codebase — no account, no card, nothing to install.

New to ADL? Read the ADL Language Reference — or see the Getting Started guide for the full workflow (API key, MCP, compile, build).

Explore the SID ecosystem

Home

SID platform overview — what this is and why it exists

Product

The full product story: compiler, memory, control plane

Documentation

ADL reference, getting started, MCP setup, architecture

ADL Reference

The language you write specs in — entities, relations, ops

API & Keys

Grab a free trial key, see endpoints, connect your harness

Live Demo

You're here — paste a spec, watch DIS compile a full app

Console

Customer App

app.sidislab.com — the customer-facing application

Evidence

Benchmarks, attestation chain, verification records

Attestations

attest.sidhq.com — verify the attestation explorer

Blog

The thinking behind SID — DIS, memory, agent architecture

Support

Help and support resources for the platform

schema.adl

meta:
  version: "2.0"
  domain: "clinic"
  @dis_generate: true
  @description: "A multi-entity clinic scheduling system with contracts, guards, and relations"

topology database: { kind: "sqlite", port: 5432 }
topology frontend: { kind: "react-app", port: 3000 }

entity Doctor:
  field id: u64 @primary_key @auto_increment
  field name: string @required
  field specialty: string @required
  field active: bool @default: true

entity Patient:
  field id: u64 @primary_key @auto_increment
  field name: string @required
  field dob: timestamp @required
  field insurance_id: string

entity Appointment:
  field id: u64 @primary_key @auto_increment
  field doctor_id: u64 @required
  field patient_id: u64 @required
  field scheduled_at: timestamp @required
  field reason: string @max_length: 512
  field status: string @default: "scheduled"

relation AppointmentDoctor:
  from: "Appointment"
  to: "Doctor"
  kind: "BELONGS_TO"
  foreign_key_field: "doctor_id"

relation AppointmentPatient:
  from: "Appointment"
  to: "Patient"
  kind: "BELONGS_TO"
  foreign_key_field: "patient_id"

contract AppointmentFutureTime:
  scope: "CreateAppointment"
  invariant: "req.scheduled_at >= now"
  enforcement: "pre_condition"
  on_violation: "abort"
  message: "Appointments must be scheduled in the future"

operation CreateAppointment:
  input: "doctor_id:u64, patient_id:u64, scheduled_at:timestamp, reason:string"
  output: "Appointment"
  create: "Appointment doctor_id=req.doctor_id, patient_id=req.patient_id, scheduled_at=req.scheduled_at, reason=req.reason"

operation CompleteAppointment:
  input: "id:u64"
  output: "Appointment"
  bindings: "appointment:Appointment@id"
  guard: "appointment.status == 'scheduled'"
  mutate: "appointment.status = 'completed'"

operation ListDoctors:
  input: ""
  output: "Doctor[]"
  query: "Doctor"

operation ListPatientAppointments:
  input: "patient_id:u64"
  output: "Appointment[]"
  query: "Appointment"

service ClinicApp:
  backend: "axum"
  database: "sqlite"
  frontend: "react-app"
  entities: "Doctor, Patient, Appointment"
  operations: "CreateAppointment, CompleteAppointment, ListDoctors, ListPatientAppointments"
  port: 8080
  health_check: "/health"

live compiler · no account · no card

Describe a system in ADL, hit Compile it, and watch the full codebase land here — backend, migrations, frontend, API, tests, Docker.

e.g. 363 lines of ADL → a 160-file clinic app, 8,116 lines

What just happened

1 · validate_adl checked syntax, semantics, topology, and authority graphs before any code existed.
2 · The compiler emitted the whole workspace: backend, database migrations, frontend, API, tests, Docker/K8s.
3 · Architecture is the source of truth — the generated code is a disposable build artifact.

Rate-limited to keep the demo fair: 5 compiles/hour per IP. Want more? Grab a free trial key — or connect Claude, Codex, or Cursor over MCP to api.sidislab.com/mcp.

Agent Lifecycle

What every agent does first.

When Claude, Codex, or Cursor connects to SID, this is the journey. Discover tools, recall patterns, read docs, report issues, see resolutions. The closed loop compounds.

Discover what's available

get_system_manifest

First thing any agent does: what tools can I call, what tiers exist, what's the platform capable of?