Contorium Runtime MCP
Runtime continuity layer MCP for AI coding agents, providing persistent workspace state and session context across tools and runs.

Contorium
Website: contorium.dev
Runtime continuity layer for AI coding agents.
AI coding tools reset context constantly.
Contorium maintains continuous workspace state across sessions, tools, and models.
It keeps AI coding agents aligned with:
- current focus
- workspace state
- active files
- git activity
- session continuity
across:
- Cursor
- VS Code
- Claude Code
- Codex
- MCP-compatible agents
⸻
What is Contorium?
Contorium defines the runtime continuity layer for AI coding agents.
Not a memory tool. Not a context retrieval system.
A persistent runtime state layer between AI agents and your workspace.
⸻
Why?
AI coding tools lose workspace state between sessions.
Every restart forces developers to:
- re-explain project goals
- rebuild architectural context
- recover debugging progress
- restore active workspace state
Contorium prevents that.
⸻
Core Runtime Capabilities
Current Focus
Continuously maintains what you’re actively building.
AI agents stay aligned with your current intent across sessions and tools.
⸻
Workspace State
Tracks the real-time state of your workspace:
- active files
- git changes
- recent activity
- working set evolution
⸻
Session Continuity
Maintains workspace continuity across:
- IDE restarts
- model switches
- long coding sessions
- multi-agent workflows
No context reset. No repeated explanations.
⸻
Quick Install
VS Code / Cursor
Install from VSIX
- Download the latest .vsix
- Open Extensions
- Click:
- ...
- Install from VSIX
- Select the downloaded file
⸻
From Source
git clone https://github.com/ContoriumLabs/contorium.git cd contorium npm install npm run compile
⸻
Quick Usage
- Set Current Focus
Define what you’re actively building.
Examples:
- Refactor authentication flow
- Fix websocket reconnect issue
- Improve MCP runtime synchronization
⸻
- Work Normally
Contorium continuously tracks:
- active files
- git activity
- workspace state
- recent changes
- session activity
No manual memory management required.
⸻
- Continue Across Sessions
Close Cursor.
Reopen tomorrow.
Your AI still understands:
- current focus
- active workspace state
- recent work
- project continuity
⸻
- Use Across AI Coding Tools
Contorium maintains runtime continuity across:
- Cursor
- VS Code
- Claude Code
- Codex
- MCP-compatible agents
⸻
MCP Integration
Contorium exposes runtime state through MCP-compatible tools.
Compatible with:
- Claude Code
- Codex
- custom MCP runtimes
- agent-based workflows
Example MCP configuration:
{ "mcpServers": { "contorium": { "command": "node", "args": ["path/to/contorium-mcp/dist/index.js"] } } }
⸻
Local-first
- no cloud sync
- no hidden telemetry
- optional BYOK
- workspace data stays local
⸻
Architecture
Contorium combines:
- IDE-native runtime tracking
- workspace state persistence
- MCP-compatible runtime access
- local-first storage
to create a continuous runtime layer for AI coding systems.
Additional docs:
- RUNTIME.md
- MCP.md
- ARCHITECTURE.md
⸻
Vision
AI coding systems need more than memory.
They need runtime continuity.
Contorium is building that layer.
Website: contorium.dev
Related Servers
Alpha Vantage MCP Server
sponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
PackageLens MCP
Lets your coding agent (such as Claude, Cursor, Copilot, Gemini or Codex) search package registries across multiple ecosystems (npm, PyPI, RubyGems, Crates.io, Packagist, Hex) and fetch package context (README, downloads, GitHub info, usage snippets)
mcp-hosts-installer
MCP server that installs and registers other MCP servers in Cursor, VS Code, or Claude Desktop from npm, PyPI, or a local folder (via npx).
Bellwether
Open-source CLI for testing MCP servers and detecting schema drift
Equibles Agent Terminal
Equibles Agent Terminal is a paid hosted remote MCP for Equibles. It exposes Streamable HTTP tool calls, bearer-token access, public server-card metadata, usage logs, and receipt-oriented JS
Agentic Control Framework (ACF)
A toolkit for autonomous agent development with tools for task management, filesystem operations, browser automation, and terminal control.
Lighthouse MCP Server
Audit web performance, accessibility, and SEO using Google Lighthouse.
MCP Terminal Server
A secure command-line interface server for the Model Context Protocol (MCP) that allows AI models to interact with a user's terminal.
MikroMCP
AI-native network automation for MikroTik RouterOS. MikroMCP exposes RouterOS as a typed, auditable Model Context Protocol server so Claude, Cursor, Codex, and other MCP clients can inspect, diagnose, and safely operate MikroTik routers in natural language.
Random Number
Provides LLMs with essential random generation abilities, built entirely on Python's standard library.
Remote MCP Server (Authless)
An example of a remote MCP server deployable on Cloudflare Workers without authentication.