Try out the MCP server and chatbot toolkit directly in a temporary Meshagent Room.

​
Prerequisites

​
Required Environment Variables

  • EXA_API_KEY=your-api-key-here   (Exa API authentication)

​
Steps

  1. Install Meshagent CLI 
    pip install "meshagent[all]"
  2. Authenticate with Meshagent Follow onboarding steps at Meshagent CLI Getting Started.
  3. Launch Exa MCP Agent in a Room 
    meshagent service test \
  --room=test \
  --role=agent \
  --image=meshagent/mcp-exa:latest \
  --env EXA_API_KEY=your-api-key-here \
  --env MESHAGENT_PORT=8001 \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-exa-service-test
    • Starts an Exa MCP server as an agent within the test Room. Room and services are temporary and will be cleaned up after inactivity.
  4. Join a Chatbot Linked to the Exa Toolkit 
    meshagent chatbot join \
  --room=test \
  --agent-name=mcp-exa \
  --name=mcp-exa \
  --toolkit=mcp-exa
    • Spawns a chatbot in the Room capable of using the Exa MCP toolkit. Multiple agents/tools can be combined in a Room.
    • The command output will include a browser link to the live Room.
  5. Test the Integration
    • Visit the provided Room URL and initiate a conversation with the mcp-exa agent to try out the Exa web search tools.

​
Project Level Deployment: Persistent MCP Tooling

For production scenarios, deploy the MCP server and chatbot as persistent project-level services. Every new Room in the project will automatically include the Exa tools and/or chatbot, removing the need to launch or manage these resources manually.

​
Persistent Agent and Chatbot Services

  1. Create the Exa MCP Agent Service 
    meshagent service create \
  --role=agent \
  --image=meshagent/mcp-exa:latest \
  --env EXA_API_KEY=your-api-key-here \
  --env MESHAGENT_PORT=8001 \
  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" \
  --name=mcp-exa-service
  2. Create a Persistent Chatbot Linked to the Exa Toolkit 
    meshagent service create \
  --image="meshagent/cli:latest" \
  --port="num=9001 path=/agent liveness=/ type=meshagent.callable participant_name=mcp-exa-chatbot" \
  --env="MESHAGENT_PORT=9001" \
  --name="mcp-exa-chatbot-service" \
  --command="meshagent chatbot service --agent-name=mcp-exa-chatbot --toolkit=mcp-exa"

​
Tools Available

​
Tools Provided

ToolShort Description
web_search_exaSearch the web using Exa AI - performs real-time web searches and can scrape content from specific URLs.

​
Tool Details

​
web_search_exa

Search the web using Exa AI - performs real-time web searches and can scrape content from specific URLs. Supports configurable result counts and returns the content from the most relevant websites.
ParameterTypeDescription
querystringSearch query
numResultsnumber optionalNumber of search results to return (default: 5)