Log Analyzer MCP
An MCP server for AI-powered log analysis that enables parsing, searching, and debugging across nine log formats directly within Claude. It features automated error extraction, sensitive data scanning, and streaming support for large log files.
README
Log Analyzer MCP
<!-- mcp-name: io.github.Fato07/log-analyzer-mcp -->
🔍 Stop copy-pasting logs into AI. Let Claude read them directly.
An MCP server for AI-powered log analysis. Parse, search, and debug log files across 9+ formats — right from Claude Code.
📊 At a Glance
| 14 MCP tools | 9+ log formats |
| 280 tests | 81%+ coverage |
🎬 Demo

Analyzing logs with 14 specialized tools
🤔 Why?
| Without log-analyzer-mcp | With log-analyzer-mcp |
|---|---|
| Copy-paste chunks of logs | Point Claude at the file |
| Lose context between pastes | Full file access |
| Manual format parsing | Auto-detection |
| Miss related errors | Smart correlation |
✨ Features
- Auto-Detection — Identifies format from 9+ common log types
- Smart Search — Pattern matching with context, regex, and time filtering
- Error Extraction — Groups similar errors, captures stack traces
- Natural Language — Ask questions like "what errors happened today?"
- Sensitive Data Scan — Detect PII, credentials, and secrets
- Multi-File Analysis — Correlate events across distributed systems
- Streaming — Handles 1GB+ files without memory issues
🚀 Quick Start
# Install (adds to Claude Code automatically)
uvx codesdevs-log-analyzer install
Then in Claude Code:
Analyze /var/log/app.log and tell me what's causing the errors
📦 Installation
One-liner (Recommended)
uvx codesdevs-log-analyzer install
Manual
<details> <summary>pip / uv / Claude Code config</summary>
# pip
pip install codesdevs-log-analyzer
# uv
uv tool install codesdevs-log-analyzer
Add to ~/.claude/settings.json:
{
"mcpServers": {
"log-analyzer": {
"command": "uvx",
"args": ["codesdevs-log-analyzer"]
}
}
}
</details>
📋 Supported Formats
| Format | Example |
|---|---|
| Syslog | Jan 15 10:30:00 hostname process[pid]: message |
| Apache/Nginx | 127.0.0.1 - - [15/Jan/2026:10:30:00] "GET /path" 200 |
| JSON Lines | {"timestamp": "...", "level": "ERROR", "message": "..."} |
| Docker | 2026-01-15T10:30:00.123Z stdout message |
| Python | 2026-01-15 10:30:00,123 - module - ERROR - message |
| Java/Log4j | 2026-01-15 10:30:00,123 ERROR [thread] class - message |
| Kubernetes | level=error msg="..." ts=2026-01-15T10:30:00Z |
| Generic | Any line with recognizable timestamp |
⚡ Performance
| Metric | Value |
|---|---|
| 100MB log file | < 10 seconds |
| Memory footprint | Streaming (no full load) |
| Max tested size | 1GB+ |
| Format detection | < 100ms |
🛠️ Available Tools
| Tool | Description |
|---|---|
log_analyzer_parse |
Detect format and extract metadata |
log_analyzer_search |
Search with context lines |
log_analyzer_extract_errors |
Extract and group errors |
log_analyzer_summarize |
Generate debugging summary |
log_analyzer_correlate |
Find related events |
log_analyzer_watch |
Monitor for new entries |
log_analyzer_ask |
Natural language queries |
log_analyzer_scan_sensitive |
Detect PII/credentials |
| + 6 more | Full reference → |
💡 Examples
Find errors:
Extract all errors from /var/log/app.log, group similar ones
Search with context:
Search for "timeout" in app.log with 5 lines of context
Correlate events:
What happened 60 seconds before each OutOfMemoryError?
Scan for secrets:
Check /var/log/app.log for accidentally logged credentials
🔧 Development
git clone https://github.com/Fato07/log-analyzer-mcp
cd log-analyzer-mcp
uv sync
uv run pytest -v --cov
📈 Star History
📄 License
MIT License - see LICENSE for details.
<p align="center"> <b>Found this useful?</b> Give it a ⭐ on GitHub!<br><br> <a href="https://github.com/Fato07/log-analyzer-mcp/issues/new?template=bug_report.yml">Report bugs</a> · <a href="https://github.com/Fato07/log-analyzer-mcp/issues/new?template=feature_request.yml">Request features</a> · <a href="https://github.com/Fato07/log-analyzer-mcp/discussions">Discussions</a> · <a href="https://github.com/Fato07/log-analyzer-mcp/blob/main/docs/TOOLS.md">Full docs</a> </p>
<p align="center"> Built by <a href="https://github.com/Fato07">Fato07</a> at <a href="https://codesdevs.io">CodesDevs</a> </p>
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。