screaming-frog-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SF_MCP_AUDIT_DIR | No | Override the directory where crawl audits are stored. Defaults to ~/.screaming-frog-mcp/audits/. | |
| SCREAMING_FROG_PATH | No | Path to the Screaming Frog SEO Spider executable if installed in a non-standard location. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| check_installA | Report whether Screaming Frog is installed, whether it is licensed, and what the current limits are. Call this first if anything else fails. |
| available_filtersA | List the export names the INSTALLED Screaming Frog build accepts. Filter names change between Screaming Frog versions and a single unknown name aborts an entire crawl, so never guess them. kind 'export-tabs' or 'save-report' contains search, e.g. 'Title' or 'Accessibility' |
| start_crawlA | Start a headless Screaming Frog crawl in the background. Returns a job_id immediately. Poll it with crawl_status. url site to crawl, e.g. https://example.com label crawl folder name (default: -) full beat the free 500-URL cap by batching sitemap URLs through list mode. Slower; use for sites over ~500 pages. everything export every report and tab filter this build supports, instead of the curated default set. Much slower. On a licensed install this also covers the API-backed tables. config path to a .seospiderconfig file. LICENSED INSTALLS ONLY: config files unlock JavaScript rendering, custom extraction and the API integrations. The free tier rejects them, so passing one here without a licence is refused up front rather than failing mid-crawl. wait_seconds block up to this long, then return the summary if the crawl finished. 0 returns straight away. |
| crawl_statusA | Check a crawl started by start_crawl. Omit job_id for the most recent. While running: elapsed time and the last runner lines. Once finished: the headline counts. Then call get_issues for the register. |
| cancel_crawlB | Stop a running crawl. Exports already written are kept. |
| list_crawlsB | List crawl folders in the audit root, newest first. |
| get_issuesA | The priority-ranked issue register from a finished crawl. This is Screaming Frog's own Issues Overview, sorted High to Low then by affected URLs. START HERE, not with raw CSV rows. crawl folder name from list_crawls, or an absolute path priority optional filter: High, Medium or Low include_fixes include the description and how-to-fix text (verbose) |
| get_analysisB | The derived analysis: what the SET of URLs means, not what each URL is. Sections: depth, link_equity, sitemap, content, performance, indexability, duplication. Each carries a 'reading' line explaining how to interpret it. section return one section only. Omit for all of them (large). |
| list_exportsA | List the CSV exports in a crawl folder, with row counts. Use this to find the right export before calling read_export. |
| read_exportA | Read rows from one CSV export, capped and filtered. crawl folder name or absolute path export file name from list_exports, e.g. 'h1_missing.csv' (the .csv suffix is optional) columns comma-separated columns to keep. Empty returns the first 8; Screaming Frog exports can be 60 columns wide. contains substring filter matched across the whole row limit max rows returned (hard ceiling 500) offset skip this many matching rows, for paging |
| build_reportA | Write a shareable write-up of a finished crawl into its folder: report.md, a self-contained printable report.html, and analysis.json. The HTML is styled, responsive, light and dark aware, and has print rules, so opening it and printing to PDF produces a clean document. There is no bundled PDF step, deliberately: shipping a headless browser to print a page your browser already prints is a bad trade. label display name in the report (default: the crawled site) |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 11 tools
Each tool targets a distinct part of the crawl workflow: installation checks, running/cancelling crawls, listing/reading exports, issue registers, derived analysis, and report building. Even the two listing tools ('list_exports' vs 'available_filters') are clearly separated by whether they refer to crawl folders or the installed build's accepted filter names.
The majority of tools follow a clear verb_noun pattern ('start_crawl', 'list_exports', 'read_export', 'cancel_crawl', 'get_issues'). 'crawl_status' and 'available_filters' are descriptive noun phrases rather than verb-led commands, creating a minor inconsistency without hurting readability.
Eleven tools is a well-scoped surface for a Screaming Frog SEO crawler. Each tool covers a distinct stage of the workflow with no redundant entries or excessive granularity.
The lifecycle is well covered: preflight checks, filter discovery, starting/status/cancelling crawls, listing crawl folders, reading exports, pulling the issue register, generating derived analysis, and building a shareable report. There are no obvious dead ends or missing operations an agent would need for the stated domain.