Skip to content
Alertly
Open navigation Close navigation

Social monitoring MCP

Ask your AI agent about your social alerts.

Connect Cursor, Claude or any MCP client to Alertly. Search curated brand mentions, competitor conversations and buyer-intent signals without writing an integration.

7-day free trial · No credit card required
An Alertly MCP search_alerts call returning a curated social monitoring alert with platform, keyword, content, intent, sentiment, relevance score, and time.
Request
search_alerts
platform: linkedin
timeframe: last_7_days
Authorization: Bearer key-xxxxxxxxxxxxxxxx
Curated alerts
{
  "platform": "linkedin",
  "keyword": "Northstar",
  "content": "Looking for a payments tool...",
  "intent": "leads",
  "sentiment": "neutral",
  "score": 9,
  "time_ago": "3 hours ago"
}

What is a social monitoring MCP server?

A social monitoring MCP server lets you search conversations, mentions and market signals from Cursor, Claude and other AI clients. Alertly filters and classifies the data before it reaches you, so your agent can work with useful alerts rather than a raw stream of posts.

01

Monitor selected sources

Alertly watches the sources and keywords configured in your account.

02

Filter and classify

Low-value matches are filtered and useful conversations are classified.

03

Ask your agent

Your MCP client searches the same curated alert stream.

Work with the conversations that matter.

Use curated Alertly alerts inside Cursor, Claude and the AI workflows your marketing, sales, research and operations teams already use.

01

Research in Cursor

Ask Cursor what people said about your brand or a competitor this week, without opening a dashboard first.

02

Weekly mention briefs

Have your agent summarize unread alerts, group them by sentiment or intent, and turn them into a short briefing.

03

Competitor recaps

Pull competitor mentions, complaints and comparison conversations into the same chat where you already write research notes.

04

Reply from buying intent

Find lead-intent posts, then draft replies or outreach from the same conversation — using the URL and context from the alert.

Useful data in. Less cleanup out.

Alertly exposes the output of your monitoring pipeline — not an unfiltered firehose of every matching post.

01

Relevance filtering

Low-value keyword matches can be filtered, leaving your agent with cleaner inputs.

02

Intent and sentiment

Use the intent, sentiment and relevance data available in your alerts to ask for the conversations that matter.

03

One normalized alert stream

Search alerts across supported sources without maintaining a separate monitoring integration for each platform.

04

Your existing monitors

The MCP server exposes alerts generated by the keywords, sources and monitoring configuration you already manage in Alertly.

MCP documentation

Start searching alerts in minutes.

Create a key, connect your MCP client to https://mcp.usealertly.com, then ask your agent to search alerts.

01

Generate an API key

Go to Account → API Keys in the app, click New, and make sure the key includes the scope alert.read.

Keys are shown masked by default — use Reveal to copy the full value. The MCP server uses the same keys as the REST API.

02

Connect your MCP client

Point Cursor, Claude or another MCP client at the Alertly server and send your API key in the header.

Server URL
https://mcp.usealertly.com
Authentication header
Authorization: Bearer key-xxxxxxxxxxxxxxxx
Cursor config (.cursor/mcp.json)
{
  "mcpServers": {
    "alertly": {
      "url": "https://mcp.usealertly.com",
      "headers": {
        "Authorization": "Bearer key-xxxxxxxxxxxxxxxx"
      }
    }
  }
}
03

Search alerts

Ask your agent to search alerts, or call search_alerts with the same filters you use in Alertly.

search_alerts arguments
Param Type Notes
platform string Comma-separated platforms (e.g. reddit,x,linkedin). Also bluesky, youtube, news.
keyword string Monitored keyword text to filter by (not a keyword ID).
offset number Pagination offset. Default: 0.
limit number Page size. Defaults to 25; values above 250 are capped at 250.
search string Searches alert title and content.
unread_only boolean Set true to only return unread alerts.
timeframe string today, yesterday, this_week, this_month, last_7_days, last_30_days, or all.
sentiment string Filter by positive, neutral, or negative. Supports comma-separated values. Use all to disable.
intent string Filter by alert intent (e.g. leads, brand_monitoring, competitor_tracking). Supports comma-separated values. Use all to disable.
Example (curl)
curl -X POST "https://mcp.usealertly.com" \
  -H "Authorization: Bearer key-xxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "search_alerts",
      "arguments": {
        "unread_only": true,
        "timeframe": "last_7_days",
        "sentiment": "positive,neutral",
        "limit": 25
      }
    }
  }'
Response shape
{
  "total": 123,
  "offset": 0,
  "limit": 25,
  "results": [
    {
      "id": "…",
      "title": "…",
      "content": "…",
      "url": "…",
      "platform": "reddit",
      "keyword": "Alertly",
      "status": "unread",
      "intent": "leads",
      "sentiment": "neutral",
      "score": 7,
      "time_ago": "3 hours ago"
    }
  ]
}

Alert content is truncated to 280 characters. Use offset to paginate when total is larger than the page.

Sources

Search alerts from every source you monitor.

Use one MCP server to access alerts generated across LinkedIn, Reddit, X, YouTube, Bluesky and Google News.

Ask Alertly from your agent.

Bring relevant social conversations into Cursor, Claude and the tools your team already uses.

7-day free trial · No credit card required

Social monitoring MCP FAQs

Answers to common questions about using Alertly alerts in Cursor, Claude and other MCP clients.

What is a social monitoring MCP server?

An MCP (Model Context Protocol) server lets AI clients like Cursor and Claude call tools. Alertly’s MCP server exposes search_alerts, so your agent can look up filtered brand mentions, competitor conversations, and related monitoring data from your account.

What is the difference between the API and MCP?

Both return the same curated Alertly alerts. The REST API is for your own software, dashboards and automations. MCP is for AI clients: connect once, then ask your agent to search alerts instead of writing fetch code.

Which clients work with Alertly MCP?

Any client that can connect to a remote MCP server with a URL and an Authorization header. Cursor is the setup we document first. Claude and other MCP clients can use the same server URL and API key.

What platforms does Alertly monitor?

Alertly currently monitors Reddit, X (Twitter), LinkedIn, Bluesky, YouTube, and Google News. You can filter search_alerts by platform when needed.

Can I filter mentions by platform or sentiment?

Yes. You can filter results by platform, keyword text, timeframe, unread status, sentiment, intent and a search string.

Can I use MCP for competitor monitoring?

Yes. You can create competitor keywords in Alertly, then ask your agent to search competitor mentions, complaints, comparison conversations, and switching signals.

Is Alertly a raw social media firehose?

No. Alertly focuses on curated social monitoring alerts rather than raw firehose access. It is designed for teams that want high-signal conversations they can review, report on, and act on.

How do I authenticate?

Create an API key in your Alertly account and include the alert.read scope. Then send it with the MCP connection using the Authorization: Bearer header. If your client offers OAuth, skip it — the key in the header is the auth.