AI doesn't remember your project.
Contorium does.

A persistent project memory layer for AI development.

Git for AI Collaboration

Scroll ↓
.contora / workspace state

Modern AI coding tools are powerful — but forget everything.

After a few interactions:

  • context resets
  • architecture gets re-explained
  • intent is lost
  • switching tools breaks understanding

You end up spending more time explaining your project than building it.

Contorium maintains a persistent cognitive layer for your project.

So AI doesn't restart from zero every time.
It continues understanding your project.

How it works

From workspace activity to any AI tool — one continuous understanding.

Workspace Activity
Project Understanding Layer
Knowledge Graph + Intent Tracking
Hotspot Detection
AI-ready Snapshot
Any AI Tool

Core features

Project cognition — not another feature checklist.

Persistent Project Memory

AI remembers your project across sessions, tools, and model switches.

Cross-tool Continuity

Cursor · Claude Code · Gemini CLI · Codex · VS Code

Project Knowledge Graph

Intent → Module → File → Function
A structured understanding of your codebase.

Hotspots

Highlights what matters right now — frequently edited files, key logic areas, and high-impact code paths.

AI-ready Snapshot

A compact project state for AI handoff.

Why it matters

Without Contorium

AI resets understanding every session.

With Contorium

AI maintains continuous project cognition.

Contorium is NOT an autonomous coding agent.

It does NOT:

  • generate tasks
  • decide what to build
  • replace developers

Developers stay in control. Contorium maintains understanding.

Install Contorium

Three peer adapters — IDE extension, MCP server, and CLI. Each reads and writes the same .contora/ state.

CLI

npm run compile
npx contorium init .
npx contorium sync .
npx contorium handoff .
npx contorium graph-snapshot .
npx contorium export .

Today's AI tools have context windows.

Tomorrow's tools need memory.

Contorium is the missing memory layer for AI development.

Because AI conversations end.
Project understanding should not.