Ask Your Project.

Understand what happened, why it changed, and what comes next — across every AI coding tool.

Powered by the Cognitive Interaction Layer (CIL), backed by a deterministic Project Intelligence Layer (PIL).

Get Started GitHub
Scroll ↓
You
"What happened this week?"
Contorium
History Decision Timeline Story

Why Contorium

AI can generate code.
But software development is about understanding.

AI forgets architecture

Structure, modules, and conventions vanish when the chat ends.

AI forgets decisions

Why JWT, why MCP, what tradeoffs you accepted — gone every session.

AI forgets project evolution

How the codebase grew and changed is never carried forward.

AI forgets why things changed

Reasoning history disappears. You re-explain instead of continuing.

Contorium preserves project intelligence so AI can reconstruct understanding instead of restarting from zero.

Ask Your Project

Natural language interface powered by CIL.

Decision Center

MCP was introduced because the team needed a standard interface for AI agents to inspect project state, transfer context, and capture structured decisions — without coupling to a single IDE or chat provider.

What happened this week?

History

Tell me everything about auth.

Knowledge Graph

What should I do next?

Action Engine

What was the project state last Friday?

Time Travel

Explain this architecture.

Story

How It Works

Every interaction becomes structured project intelligence that can be queried later.

Capture
Structure
Cognitive Events
Ask
Answer

The Intelligence Stack

CIL on top. PIL underneath.

You
CIL
Kernel
PIL
.contora

CIL understands. PIL remembers.

What Contorium Stores

Project State Architecture Decisions Project History Knowledge Graph Snapshots Impact Health Timeline

Everything stays inside your repository.

Built for Humans and AI

One Cognitive Kernel — three ways to ask, capture, and explore.

Developer
CIL
AI Agent

IDE

Ask Contorium while you code.

  • Ask · History · Decisions
  • Capture focus & notes
  • Project visualization

MCP

CIL for AI agents.

  • ask_project
  • Transfer context & story
  • Cognitive health

CLI

Ask from the terminal.

  • contorium ask
  • History · DNA · health
  • Inspect & transfer

Works Everywhere

One project. One intelligence layer. Every runtime.

"Why not Redux?"

Decision Center returns the ADR, alternatives considered, and reasoning — not a guess from chat memory.

"What do we know about auth?"

Knowledge Graph connects modules, decisions, and timeline events around a single entity.

Switch Cursor → Claude Code

Transfer project story or context — continue without re-explaining architecture.

CIL Features

Built for long-lived AI-assisted projects.

Project History

Unified timeline of cognitive events

Decision Center

ADR-style records with Why and alternatives

Time Travel

Snapshot replay for any date

Knowledge Graph

Everything related to an entity

Cognitive Health

Missing WHY, stale decisions, conflicts

Project Story

Goal + journey narrative for onboarding

Project DNA

Identity fingerprint for agent handoff

Evolution Journey

How the project grew over time

Local First

No Cloud No Vendor Lock-in Deterministic Git Friendly Own Your Intelligence

Open Architecture

Everything routes through one Cognitive Kernel.

CIL
Kernel
PIL
.contora
IDE MCP CLI Dashboard

Open Source

MIT License — inspect, fork, and run entirely on your machine.

GitHub ContoriumLabs/contorium npm @contorium/mcp MCP Setup for Cursor, Claude, Codex VS Code IDE extension · Marketplace

Build AI that understands projects.

Contorium lets you ask your project what happened, why it changed, and what comes next.

GitHub Install Documentation

Local-first · MIT License · CIL v3