EN
BosanskiDeutschEnglish
Log in Download free
Guide · chapter 08

Command line and MCP server

Run crawls and exports without the app window, from a script or Task Scheduler, or let an AI agent read the findings over MCP.

2 min readUpdated 21 October 2026

Everything you do in the app you can also do without it: from a script, from Task Scheduler or through an AI agent. The settings come from the same presets.

Command line

seo-bee crawl https://example.com --config shop --export response_codes:internal_4xx --out D:\reports
seo-bee crawl --list urls.txt --report redirect_chains --out D:\migration
seo-bee export <crawl-id> --bulk all_inlinks --format xlsx

You refer to tabs, filters and reports by ID: seo-bee reports lists every report and bulk export, seo-bee list every saved crawl, seo-bee presets the presets.

CommandWhat it does
crawl <url>Crawls the whole site. With --list file it checks only the addresses in the file.
--configName of the settings preset.
--export, --report, --bulkTab and filter (tab:filter), report or bulk export. You can give several, separated by commas.
--render, --max-urls, --depthJavaScript rendering and limits, when you're not using a preset.
--outFolder for the exported files.
export <crawl>Exports from a crawl that is already saved.
serveStarts the app and its schedules without opening a browser.
mcpStarts the MCP server for AI agents.

Windows Task Scheduler

To crawl while the app is closed, create a task in Task Scheduler that runs seo-bee crawl with your preset. The exports end up in the folder, and the crawl is saved and shows up in the app like any other.

MCP server for AI agents

The seo-bee mcp command starts a server that speaks the Model Context Protocol. An AI agent in Claude Code, Claude Desktop or another MCP client can then start a crawl by itself and read the findings. The tools the agent gets:

  • start_crawl, crawl_status, control_crawl, list_crawls
  • list_tabs, get_table, get_url, list_issues
  • list_reports, get_report, export_file, compare_crawls
  • map_redirects, site_structure, list_segments, list_presets
  • pull_search_console, pull_google_analytics

Example: "Send the bee to example.com, then give me every page with a duplicate title and suggest new ones." The agent starts the crawl, waits for it to finish, reads the Duplicate filter in the Page titles tab and replies with a table.

That's the end of the guide. For specific tasks, go to the tutorials.