Frequently Asked Questions

Answers to common questions about the SID Platform

General Questions

What is SID?: SID (Singular Intelligence Domain) is a Constitutional Operating System for AI agents that enables autonomous application development through Architecture Description Language (ADL) compilation.
How does SID differ from traditional development?: With SID, you describe what you want in ADL (a high-level architectural specification), and the DIS compiler generates a complete, production-ready application including backend, frontend, database schema, tests, and deployment configurations. You never write boilerplate code.
What is ADL?: ADL (Architecture Description Language) is a domain-specific language for specifying software architecture. It's human-readable, indentation-sensitive, and focuses on declaring components, relationships, and behaviors rather than implementation details.

What technologies does SID generate?: DIS compiles ADL into a backend (Rust Axum, Node/TypeScript Express, or Python FastAPI), a frontend (React, Vue, Svelte, Solid, or Leptos), database schema and migrations (PostgreSQL or SQLite), API docs, tests, and Docker/Kubernetes deployment setup. Every compiled run is hashed into the SHA-256 ledger — see the evidence page for live hashes. Unsupported backend kinds (actix, hono, fastify, go-gin) are rejected loudly at validation rather than emitting the wrong server. The engine also emits C++ for its internal system/config domains, but those are not application backends.
Can I customize the generated code?: Yes. Two mechanisms keep your manual work through regeneration: wrap code in BEGIN/END MANUAL markers inside generated files (snapshotted and re-injected on every build), or place whole files under custom/ which are copied over the generated tree after each regeneration. Everything else stays in sync with the spec.
What is the Model Context Protocol (MCP)?: MCP is a standardized way for AI assistants to interact with external tools and resources. SID exposes its functionality through an MCP server, allowing AI agents like Claude Code, Cursor, and others to directly invoke DIS compilation, access memory, and manage workflows.

How do I get access to SID?: Visit sidai.live to sign up for access. Connect your AI app in one click with OAuth (paste https://api.sidislab.com/mcp into Add Custom Connector, approve in your browser), or use an API key from your dashboard.
What's the difference between the trial and developer tiers?: The trial tier offers 1,000 calls at 100 requests/minute. The developer tier offers 10,000 calls at 100 requests/minute. Both expose the same 30 external MCP tools (see the categorized API reference). The internal mission tools are not on the external surface.
What are Asset Passports?: A signed proof of what an asset is. verify_repo hashes every file (SHA-256), folds the hashes into a Merkle root, and signs it with an HMAC — an AttestationV1. get_passport reads the attestation, verify_passport recomputes the root and signature to confirm nothing was tampered with, passport_summary gives a one-line thumbnail, and list_passports shows what you minted. Source stays sealed — only fingerprints, structure, and build/test outcomes are exposed. This is how agents and deal-rooms trust a repo without reading it.
Where can I find examples of ADL?: Examples are available in the API Reference section and in the Tutorials section.