elmah.io MCP のリモート MCP サーバーとは何ですか?
elmah.io MCP のリモート MCP サーバーは、https://mcp.elmah.io/mcp にホストされた Model Context Protocol エンドポイントです。AI アシスタントはローカルに何もインストール・実行することなく接続できます。

@elmah.io
基本情報
The elmah.io MCP server allows AI assistants to interact directly with your error logs, deployments, and organization statistics. Debug production issues faster than ever before.
https://mcp.elmah.io/mcpclaude mcp add elmah-io-mcp --transport http https://mcp.elmah.io/mcpツール一覧はまだありません
この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。
このサーバーは認証が必要です。ツールの取得にはログインや提供元の認証情報が必要な場合があります。
上のボタンでサーバーのエンドポイントに接続し、ツール一覧を取得できます。
Error logging and uptime monitoring for AI assistants. Connect your AI tools directly to elmah.io to search logs, investigate errors, check deployments, and monitor uptime, all through natural language.
Server URL: https://mcp.elmah.io/mcp
Transport: Streamable HTTP
Auth: OAuth2 (interactive clients) or API key Bearer token (scripts, CI/CD)
Status: Beta
elmah.io is a cloud-based error logging and uptime monitoring service for .NET and other platforms. Applications send errors and heartbeats to elmah.io, where they're searchable, groupable, and alertable. The MCP server gives AI assistants read access to that data, plus a few targeted write actions (hide, fix, reopen, and mark messages as bot traffic), so you can debug production issues without leaving your editor or chat client.
Ask "show me the 5 most recent errors in my Production log" or "what changed since my last deployment," and your assistant pulls the answer straight from live data instead of you digging through a dashboard. It can fetch a full stack trace by ID, so there's no copy-pasting between tabs, and group errors by frequency to surface which bugs are actually worth fixing. Deployments and uptime checks are queryable the same way: list recent releases with their error counts, or check whether a monitored endpoint is currently up. Authenticate with an API key instead of OAuth and every call runs against your organization's full set of logs, not just what one user account can see.
| Tool | Description |
|---|---|
users_get_current | Fetch the current user details. |
system_get_context | Get system information like current date and time. Useful when calculating time intervals. |
organizations_list | List all organizations the current user belongs to (or the current organization, when authenticated with an API key). |
organizations_get_details | Get detailed organization info: configuration, subscription, and usage statistics. |
logs_list | List logs the current user has access to. |
logs_get_details | Get detailed configuration and metadata for a specific log. |
messages_get | Fetch a single log message by ID, including stack trace and metadata. |
messages_list_recent | List the most recent log messages for a log. |
messages_list_frequent | Find the most frequent/common error groups in a log, useful for spotting noisy bugs. |
messages_count | Count log messages by severity and timeframe. |
messages_hide / messages_unhide | Hide or unhide a log message by ID (optionally, every message sharing its grouping hash). |
messages_fix / messages_open | Mark a log message as fixed, or reopen it. |
messages_bot / messages_unbot | Mark or unmark log messages as bot-generated traffic. |
deployments_list | List recent deployments for a log, optionally including error metrics per deployment. |
uptime_list | List uptime checks for a log with current status and 24h up/down percentage. |
uptime_get_details | Get real-time results and regional performance metrics for an uptime check. |
heartbeats_list | List heartbeats configured for a log with current status. |
heartbeats_get_details | Get detailed check-in history for a heartbeat. |
recommendations_list | List automated recommendations for improving error monitoring setup, security, and configuration. |
recommendations_get_details | Get the full description and suggestions for a single recommendation. |
Destructive tools (hide, fix, reopen, unhide, bot marking) only run when a user explicitly asks. The tool descriptions themselves instruct the assistant not to call them speculatively while just investigating.
Add the server to your client using the streamable HTTP URL:
Most clients (Claude Desktop, Claude Code, VS Code, Cursor, Visual Studio, ChatGPT) walk you through an OAuth2 sign-in the first time you connect. Tokens are valid for 30 days.
Claude Code, from the terminal:
claude mcp add --transport http --client-id claudecode elmahio https://mcp.elmah.io/mcp
Then run /mcp inside Claude Code and authenticate.
Scripted / CI access: pass an elmah.io API key as a Bearer token instead of going through OAuth. Requests then run in the context of your organization rather than a specific user.
curl -s -X POST https://mcp.elmah.io/mcp \
-H "Authorization: Bearer API_KEY" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"logs_list","arguments":{}}}'
Full setup guides for each client: https://docs.elmah.io/setup-mcp-server/
elmah.io AI Plugin
For Claude Code and Cursor, the elmah.io AI Plugin (https://github.com/elmahio/elmah-io-ai-plugin) (Apache 2.0) packages this MCP server together with prebuilt skills (debug-production-error, post-deployment-check, investigate-frequent-errors, check-uptime-status) and a background deployment-watchdog agent that watches for regressions after a deploy.
claude plugin marketplace add elmahio/elmah-io-ai-plugin claude plugin install elmah-io@elmah-io
Links
elmah.io MCP のリモート MCP サーバーは、https://mcp.elmah.io/mcp にホストされた Model Context Protocol エンドポイントです。AI アシスタントはローカルに何もインストール・実行することなく接続できます。
エンドポイント https://mcp.elmah.io/mcp を Claude Code、Cursor、VS Code など任意の MCP 対応クライアントに追加してください。このページのセットアップ手順で各クライアントを一度に設定できます。
はい。elmah.io MCP は OAuth を使用します。初回接続時に MCP クライアントがブラウザを開いてサインイン・認可を行い、以降はその認証情報を再利用します。
elmah.io MCP は Streamable HTTP エンドポイントを公開しており、これはリモート MCP サーバーで広く使われ、主要な MCP クライアントがサポートするトランスポートです。
Understand and troubleshoot your Journeys and Campaigns
Convert web pages to clean Markdown and retrieval-ready chunks with browser rendering support.
Marketing campaign and audience insights from Adobe
Search and read current AWS documentation, check regional availability and load AWS agent skills, with no AWS account needed.
Write, schedule and publish LinkedIn posts in your own voice, reply to comments and read your analytics, all from your AI assistant. Remote MCP server with OAuth sign-in.
コメント