Deployment

Chicory agents can be deployed through two main methods: REST API for direct API access and MCP Gateway for integration with MCP-compatible systems. (A2A coming soon)

Option 1: Deploy as REST API

This method provides direct API access to your agent for integration with applications and services.

Steps for REST API Deployment

1. Access Agent Configuration

   • Click the Agent's Name at the top of the chat interface

   • Click "Edit" to open the agent's configuration

2. Deploy the Agent

   • Navigate to the section labeled "Deployed API"

   • Click "Deploy Agent"

3. Get Deployment Details

   • After deployment, you'll see the API endpoint, authentication details, and version info

   • Generate an API token from Chicory platform

   • Store it in your target repo secrets as CHICORY_API_TOKEN

   • Similarly, store your agent ID as CHICORY_AGENT_ID

Using the REST API

Once deployed, you can interact with your agent via HTTP requests:

   curl -X POST https://app.chicory.ai/api/v1/runs \
   -H "Content-Type: application/json" \
   -H "Authorization: Bearer ${CHICORY_API_TOKEN}" \
   -d '{
   "agent_name": "${AGENT_ID}",
   "input": [{
      "parts": [{
         "content_type": "text/plain",
         "content": "Your question here"
      }],
      "created_at": "2025-09-10T16:38:27.056Z"
   }]
   }'

Option 2: Deploy as MCP Gateway

MCP Gateways enable you to create a group of agents and execute complex agent workflows. Each agent is deployed to an MCP gateway in an agentic process, allowing for orchestrated multi-agent interactions and collaborative task execution. Refer Documentation.

Steps for MCP Gateway Creation

1. Create MCP Gateway

   • Navigate to the MCP Gateway section in your dashboard

   • Click "Create New MCP Gateway"

   • Name: Enter a descriptive name for your gateway (e.g., "Data Analysis Workflow")

   • Description: Describe the purpose and workflow of your agent group (e.g., "Orchestrates data extraction, analysis, and reporting agents")

   
2. Deploy Agents to MCP Gateway

   • Select the agents you want to include in the workflow as a tool

   • For each tool, valdiate/configure the description

   

MCP Gateway Benefits

• Multi-Agent Orchestration: Coordinate multiple specialized tools (agents) in complex workflows

• Agentic Process Execution: Enable agents to work together autonomously on multi-step tasks

• Scalable Architecture: Handle complex tasks by distributing work across specialized agents

• Process Standardization: Create reusable agent workflows for common business processes

Deployment Best Practices

• Security: Always use secure token storage and rotation practices

• Monitoring: Set up monitoring for both API and MCP deployments

• Testing: Thoroughly test deployments in staging before production

• Documentation: Document your API endpoints and MCP configurations for your team

Next Steps

After deployment, consider:

• Setting up monitoring and logging

• Implementing rate limiting and security measures

• Creating documentation for API consumers

• Planning for scaling and load balancing

---