MCP Integration Guide

Integrate APEX with your AI agent

APEX provides a Model Context Protocol (MCP) server that lets AI agents discover, search, and analyze AI news programmatically.

Status: public MCP package publication is in progress. This guide documents the target integration shape; public install commands will be finalized when the repository/package is published.

What is MCP?

MCP (Model Context Protocol) is a standard interface that lets AI models access external tools and data sources. APEX's MCP server exposes search, topic following, and daily briefing capabilities.

Installation

Public package command will be published with the MCP repository.

Configuration

Claude Code (Desktop)

1. Open Claude Code settings → MCP Servers
2. Add new server:
   • Name: apex
   • Command: <mcp-command> (to be published with MCP package)
   • Env: APEX_API_KEY=your-key (optional for public endpoint)

Alternatively, add to your claude_config.json:

{
  "mcpServers": {
    "apex": {
      "command": "python3",
      "args": ["-m", "apex_mcp.server"], 
      "env": {
        "APEX_API_KEY": "your-key-here"
      }
    }
  }
}

OpenCode

1. Create .opencode/config.json in your project:

{
  "mcp": {
    "apex": {
      "command": "python3",
      "args": ["-m", "apex_mcp.server"]
    }
  }
}

2. Restart OpenCode and connect to apex server.

Cursor

1. Go to Settings → MCP → Add Server
2. Configure:

   Name: APEX Intelligence
   Command: <mcp-command>
   Args: --api-key=your-key
   

Available Tools

Once connected, your agent can use these MCP tools:

1. search_news

Search APEX news database.

Parameters:

• query (optional): Free-text search query
• topics (optional): Array of topic filters
• from_date (optional): ISO date string
• to_date (optional): ISO date string
• limit (optional): Max results (default: 20)

Example usage in agent:

Search for AI news about "healthcare" published in the last week.

Returns: Array of news items with title, summary, source, topics.

2. follow_topics

Subscribe to topic feeds for briefings.

Parameters:

• topics: Array of topic strings to follow

Example:

Follow these topics: ai, healthcare, finance

3. get_daily_briefing

Get a daily briefing on topics.

Parameters:

• topics (optional): Override followed topics
• date (optional): Date for briefing (default: today)

Example:

Get me today's AI news briefing

Workflow Examples

Example 1: Automated News Research Agent

# Agent prompt
"""
You are an AI research assistant. Use the APEX MCP tools to help the user stay updated.

When the user asks about AI news:
1. Use `search_news` with their query
2. If they want ongoing updates, use `follow_topics`
3. Summarize findings in a clear, actionable format
"""

Example 2: Daily Briefing Automation

# Agent prompt
"""
Every morning, check APEX for a daily briefing:
1. Call `get_daily_briefing` for topics: ai, business, regulation
2. Summarize the top 5 items
3. Flag any regulatory or compliance alerts
"""

Example 3: Sector-Specific Intelligence

# Agent prompt
"""
The user is a healthcare startup founder. Configure your APEX integration:

1. Follow topics: healthcare, ai-regulation, medical-devices
2. When asked about competition, search news for competitor names
3. Weekly: generate a sector intelligence report from briefing results
"""

Rate Limits & Quotas

• Free tier: 100 API calls/day
• Pro tier ($19/mo): 1000 calls/day
• Team ($39/mo): 5000 calls/day, team sharing

Monitor usage via GET /health endpoint.

Troubleshooting

"MCP server not found"

Ensure the server process is running. Check your MCP client's process manager.

"Authentication required"

Set APEX_API_KEY environment variable or pass --api-key in server args.

"Topic not recognized"

APEX supports standard topic taxonomy: ai, llm, healthcare, finance, regulation, ml, research, etc. Use GET /topics to list valid topics.

Full Example: Claude Code Integration

1. Install MCP server Use the public package command once MCP repository/package is published.

2. Configure Claude Code (~/.config/claude/settings.json)

   {
     "mcpServers": {
       "apex": {
         "command": "python3",
         "args": ["-m", "apex_mcp.server"],
         "env": {
           "APEX_API_KEY": "sk_test_..."
         }
       }
     }
   }
   

3. Use in conversation

   User: What's new in AI regulation this week?
   
   Claude: Let me search APEX for recent regulatory news...
   [Uses `search_news` with query="regulation", from_date="2025-11-24"]
   [Presents 3-5 relevant items with summaries]
   

Next Steps

• View full API reference: https://apex.incwo.com/docs/api-reference
• CLI status: in progress (incwo/apex-cli)
• MCP package status: publication in progress