Open Multi Agent Canvas

v1.0.0Coding Agentsstable

The open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add MCP servers for deep research

ai-agentscopilotkitmcp-clientmulti-agentopen-canvas
Share:
493
Stars
0
Downloads
0
Weekly
0/5

What is Open Multi Agent Canvas?

Open Multi Agent Canvas is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add mcp servers for deep research

The open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add MCP servers for deep research

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • The open-source multi-agent chat interface that lets you man

Use Cases

Manage multiple AI agents in one dynamic conversation interface.
Add MCP servers for deep research and enhanced agent capabilities.
CopilotKit

Maintainer

LicenseMIT
Languagetypescript
Versionv1.0.0
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx open-multi-agent-canvas

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Open Multi Agent Canvas

Open Multi-Agent Canvas is an open-source, browser-based chat interface built with CopilotKit that lets users orchestrate multiple AI agents — including travel planners, research agents, and general-purpose assistants — within a single dynamic conversation. Users can extend the canvas by connecting any MCP server through the built-in configuration panel, enabling deep research and tool-augmented agent capabilities using Standard IO or SSE transport. Developers building multi-agent applications or researchers who want a flexible UI for coordinating specialized AI agents across complex workflows will find this canvas a strong foundation.

Prerequisites

  • Node.js 18 or later and pnpm installed
  • Python 3.10 or later with Poetry (for the optional agent backend)
  • A CopilotKit Cloud account and public API key (NEXT_PUBLIC_CPK_PUBLIC_API_KEY)
  • An OpenAI API key (OPENAI_API_KEY) for the LangGraph agent backend
  • Optional: LangSmith API key (LANGSMITH_API_KEY) for agent tracing
1

Clone the repository

Clone the Open Multi-Agent Canvas repository to your local machine.

git clone https://github.com/CopilotKit/open-multi-agent-canvas
cd open-multi-agent-canvas
2

Configure frontend environment variables

Create a .env file in the frontend directory with your CopilotKit Cloud public API key.

cd frontend
cp .env.example .env
# Edit .env and set:
# NEXT_PUBLIC_CPK_PUBLIC_API_KEY=your_copilotkit_key
3

Install frontend dependencies and start the app

Install pnpm dependencies and build and start the Next.js frontend.

pnpm i
pnpm run build && pnpm run start
4

Start the optional MCP agent backend

For the MCP-enabled research agent, start the LangGraph backend in the agent directory with your OpenAI credentials.

cd ../agent
cp .env.example .env
# Set OPENAI_API_KEY and LANGSMITH_API_KEY in .env
poetry install
poetry run langgraph dev --host localhost --port 8123 --no-browser
5

Add MCP servers through the UI

In the running canvas app, click the 'MCP Servers' button in the top right. Add servers via Standard IO (local scripts) or SSE (remote MCP endpoints like mcp.composio.dev or mcp.run).

Open Multi Agent Canvas Examples

Client configuration

Frontend .env file configuration for Open Multi-Agent Canvas.

{
  "mcpServers": {
    "open-multi-agent-canvas": {
      "command": "npx",
      "args": ["open-multi-agent-canvas"],
      "env": {
        "NEXT_PUBLIC_CPK_PUBLIC_API_KEY": "your_copilotkit_key",
        "OPENAI_API_KEY": "your_openai_key"
      }
    }
  }
}

Prompts to try

Example tasks to try in the multi-agent canvas once running.

- "Research the latest developments in MCP server tooling and summarize the key trends"
- "Plan a 5-day trip to Tokyo including flights, hotels, and daily itinerary"
- "Use the research agent to find three academic papers on multi-agent coordination"
- "Switch to the travel agent and find the cheapest flights from NYC to London next month"
- "Add the mcp.composio.dev GitHub server and then list my open pull requests"

Troubleshooting Open Multi Agent Canvas

Frontend fails to start with 'NEXT_PUBLIC_CPK_PUBLIC_API_KEY is not defined'

Ensure you have created the .env file inside the frontend/ directory (not the repo root) and set NEXT_PUBLIC_CPK_PUBLIC_API_KEY to a valid key from your CopilotKit Cloud dashboard.

LangGraph agent backend fails to connect or agents do not respond

Verify the agent backend is running on port 8123 with 'poetry run langgraph dev --host localhost --port 8123 --no-browser'. Check that OPENAI_API_KEY is set correctly in the agent/.env file.

Adding an MCP server via SSE fails or shows connection error

Ensure the remote MCP server URL is publicly accessible and supports SSE transport. For local servers using Standard IO, verify the command and script path are correct and the script is executable.

Frequently Asked Questions about Open Multi Agent Canvas

What is Open Multi Agent Canvas?

Open Multi Agent Canvas is a Model Context Protocol (MCP) server that open-source multi-agent chat interface that lets you manage multiple agents in one dynamic conversation and add mcp servers for deep research It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Open Multi Agent Canvas?

Follow the installation instructions on the Open Multi Agent Canvas GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Open Multi Agent Canvas?

Open Multi Agent Canvas works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Open Multi Agent Canvas free to use?

Yes, Open Multi Agent Canvas is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "open-multi-agent-canvas": { "command": "npx", "args": ["-y", "open-multi-agent-canvas"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Open Multi Agent Canvas?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides