System Architecture

One brain.
Five tools. Zero drift.

This page explains how Spiral works under the hood. If you just want to use it, start here instead.

Spiral uses a central facts database as your project's single source of truth. Every IDE configuration file, every context document, every tool-specific rule is automatically generated from this central brain. Four passive sync layers keep everything current without a single manual command.

The Architecture

How knowledge flows from the central brain to every tool.

Central Facts DB
Context Server
Cursor
Claude Code
GitHub Copilot
Windsurf
Other Tools

Hover over any node to see its connection to the DB

Capabilities

Everything your AI tools need.

Structured Memory Database

A vector-indexed facts database is your project's single source of truth. Every context file your tools use is a read-only projection of this central brain.

pgvector

Knowledge Graph

Every fact links to related facts. See your project as a connected, navigable graph — features connect to constraints connect to decisions.

∞ links

Curated Skill Library

Thousands of community-built AI guardrails, pre-matched to your stack and safety-scored to filter out harmful patterns.

66K+

Instant Scaffolding

Describe your idea. Get a complete repo with real code, not placeholder boilerplate.

~30s

Universal Context Sync

Cursor, Claude Code, Copilot, Windsurf, and more — all stay current via real-time push and file system watching.

5 tools

Human-in-the-Loop

Every AI suggestion shows a diff and rationale. You approve before anything changes.

100%
Skill Registry

7 skills that actually protect you.

Included in every project. Built for developers who ship fast and need guardrails they didn't know to ask for.

spiral-sync

Prevents semantic drift. Monitors changes across IDEs and triggers a pipeline extraction to update the central facts database.

Included in every Spiral project

Context Sync

Four layers keep everything current.

Zero commands to remember. Context stays in sync passively.

Real-Time Context Push

Push

Spiral pushes live updates to every subscribed IDE tool using the Model Context Protocol. No polling, no manual refresh.

Cross-Device Sync

Multi-device

Mirrors your knowledge graph across all devices. Edit on your phone, sync to your workstation. Always in sync, everywhere.

Auto-regeneration

Derived

Tool-specific config files auto-regenerate whenever the core database updates. Cursor rules, Claude context, Copilot instructions — all derived automatically.

Team Broadcast

Multiplayer

When a teammate records a decision, your local knowledge updates in the background. No merge conflicts, just shared context.

See the architecture in action.

Describe your project and watch Spiral build the knowledge graph in real time. Free to start.

Get Started