LOGAN is a fast, AI-ready log file viewer built on Electron that also exposes an MCP-compatible interface for AI agents and external tooling. It supports powerful log analysis, filtering, and visualization features, plus a built-in agent and an HTTP API for custom integrations. With MCP, you can connect Claude Code, Cursor, or other MCP clients to discover and invoke LOGAN's analysis tools (pattern detection, context search, time-range analysis, crash investigation, baseline comparisons, and more) directly from your agent workflows. The app also offers a live data surface through Serial, Logcat, SSH, and other connections, letting agents correlate static logs with live streams in real time.

To use LOGAN via MCP, install and run LOGAN as described in the installation section, then connect an MCP client to the running LOGAN instance. You can discover and invoke tools exposed by LOGAN through the standard MCP tooling model, open a project, and leverage features like context search, time-range analysis, and bookmarks from within your agent workflows. If you prefer programmatic access, you can utilize the built-in HTTP API to perform log analyses, start agent sessions, or fetch metadata about files and analyses.

Prerequisites:

• Node.js 18+ (recommended) and npm/yarn installed
• Optional: ripgrep installed for fastest search performance

Install from source (recommended):

1. Clone the repository git clone https://github.com/SolidKeyAB/logan.git
2. Navigate to the project cd logan
3. Install dependencies npm install
4. Run the application (development mode) npm start

   This builds and launches LOGAN

Packaging / distribution (optional): 5) Create distributable bundles for all platforms npm run package # macOS npm run package:win # Windows npm run package:linux # Linux npm run package:all # all platforms

CLI and agent integration setup: 6) Register the CLI tool globally (if you plan to invoke LOGAN from terminal) npm link 7) Run LOGAN with a log file or directory as an argument from the CLI logan myfile.log

Notes:

• Ensure your environment has network access if you plan to use the HTTP API or MCP integrations.
• If ripgrep is not installed, LOGAN will fall back to a stream-based search.

Tips and common considerations:

• MCP integration: LOGAN exposes a suite of log-analysis and navigation tools via MCP. Ensure your MCP client is configured to discover and connect to the LOGAN instance, typically using the HTTP API path exposed after startup.
• Live connections: LOGAN can attach up to four live streams (Serial, Logcat, SSH, Datadog). Check that the target devices or endpoints are accessible and that authentication (keys, passwords) is configured as needed.
• Per-file and global storage: LOGAN persists bookmarks, highlights, notes, and history in ~/.logan and per-file .logan folders. Back up these storage locations if you rely on long-term analysis.
• Performance: For very large logs, enable virtual scrolling and use ripgrep for fast searches. If ripgrep is not available, LOGAN will still function with a fallback search but may be slower.
• Troubleshooting: If the MCP integration does not appear in your client, verify that the LOGAN HTTP API is reachable and that the correct endpoints are exposed. Review logs for any authentication or CORS issues when accessing from a different host.