# Frontshelf > Frontshelf crawls your production site, finds the highest-impact technical defects, and returns the evidence your coding agent needs to fix them, then independently verifies the fix landed. Frontshelf is the technical SEO layer for the agent era: a cloud crawler and audit engine whose users are both founders and their AI coding agents (via MCP server, CLI and REST API). The customer's own agent fixes defects; Frontshelf's independent verifier confirms the fix landed in production. This file is generated from the site's content graph, the same source as its sitemaps. llms.txt is a proposed convention, not a ranking signal; it is offered as a convenience for agents and LLMs. ## Docs - [Agent keys](https://frontshelf.io/docs/agent-keys): Scoped personal access tokens for agents: the scopes, per-site restriction, expiry, the reveal-once secret, and the Keys section of Agents that makes them. - [REST API](https://frontshelf.io/docs/api): Every HTTP route: Agent-key auth, the scope each route needs, the deletes' optional Idempotency-Key, untrusted-content blocks, error codes, and the public signed doors. - [FrontshelfBot](https://frontshelf.io/docs/bot): The crawler's identity: its user agent, how to verify a hit by reverse DNS or the published IP list, how fast it crawls, and how to block it in robots.txt. - [articles_get](https://frontshelf.io/docs/capabilities/articles_get): One finished article as it will publish, with its images - [articles_list](https://frontshelf.io/docs/capabilities/articles_list): The site's finished articles, ready to publish - [articles_report_live](https://frontshelf.io/docs/capabilities/articles_report_live): Say where a finished article is live on the site - [attempts_get](https://frontshelf.io/docs/capabilities/attempts_get): One repair attempt: its work order, steps, actions and verdicts - [changes_get](https://frontshelf.io/docs/capabilities/changes_get): What changed between two crawls of a site - [comparisons_get](https://frontshelf.io/docs/capabilities/comparisons_get): Two crawls of a site compared, filter by filter - [comparisons_start](https://frontshelf.io/docs/capabilities/comparisons_start): Build the comparison of two crawls - [configs_get](https://frontshelf.io/docs/capabilities/configs_get): Read a site's crawl settings, asked and effective - [configs_update](https://frontshelf.io/docs/capabilities/configs_update): Change a site's crawl settings - [crawls_get](https://frontshelf.io/docs/capabilities/crawls_get): One crawl's status, progress and what it left out - [crawls_start](https://frontshelf.io/docs/capabilities/crawls_start): Start a crawl of a site - [elements_list](https://frontshelf.io/docs/capabilities/elements_list): The elements of one finding: the things on that page the rule is about - [filters_list](https://frontshelf.io/docs/capabilities/filters_list): List the Explorer's tabs and every filter on them, the names pages_list filters by - [findings_ignore](https://frontshelf.io/docs/capabilities/findings_ignore): Set some of an issue's findings aside - [findings_list](https://frontshelf.io/docs/capabilities/findings_list): The findings of one issue, one row per page - [findings_unignore](https://frontshelf.io/docs/capabilities/findings_unignore): Put set-aside findings back in the open count - [gsc_alerts_disable](https://frontshelf.io/docs/capabilities/gsc_alerts_disable): Turn a site's weekly drop check off - [gsc_alerts_enable](https://frontshelf.io/docs/capabilities/gsc_alerts_enable): Turn a site's weekly drop check on - [gsc_alerts_list](https://frontshelf.io/docs/capabilities/gsc_alerts_list): A site's Search Console drop alerts - [gsc_analytics_get](https://frontshelf.io/docs/capabilities/gsc_analytics_get): Google Analytics organic and AI traffic beside Search Console - [gsc_analytics_pages_list](https://frontshelf.io/docs/capabilities/gsc_analytics_pages_list): Google Analytics landing pages for one audience - [gsc_brand_set](https://frontshelf.io/docs/capabilities/gsc_brand_set): Set a site's Search Console brand terms - [gsc_changes_list](https://frontshelf.io/docs/capabilities/gsc_changes_list): The page-change markers of a site's Search chart - [gsc_compare_get](https://frontshelf.io/docs/capabilities/gsc_compare_get): A site's Search Console rows with two filters side by side - [gsc_counts_get](https://frontshelf.io/docs/capabilities/gsc_counts_get): A site's distinct Search Console queries per day - [gsc_filters_delete](https://frontshelf.io/docs/capabilities/gsc_filters_delete): Delete a saved Search Console filter - [gsc_filters_list](https://frontshelf.io/docs/capabilities/gsc_filters_list): The saved Search Console filters a site offers - [gsc_filters_save](https://frontshelf.io/docs/capabilities/gsc_filters_save): Save a Search Console filter - [gsc_index_get](https://frontshelf.io/docs/capabilities/gsc_index_get): Index monitoring for a site - [gsc_index_movements_list](https://frontshelf.io/docs/capabilities/gsc_index_movements_list): The days Google's answer about a tracked page changed - [gsc_index_page_get](https://frontshelf.io/docs/capabilities/gsc_index_page_get): Google's index history of one page - [gsc_index_pages_list](https://frontshelf.io/docs/capabilities/gsc_index_pages_list): The tracked pages and Google's latest answer about each - [gsc_links_list](https://frontshelf.io/docs/capabilities/gsc_links_list): A site's client links - [gsc_links_revoke](https://frontshelf.io/docs/capabilities/gsc_links_revoke): Revoke a client link - [gsc_links_save](https://frontshelf.io/docs/capabilities/gsc_links_save): Make a client link, or save the next version of one - [gsc_notes_delete](https://frontshelf.io/docs/capabilities/gsc_notes_delete): Delete a Search Console note - [gsc_notes_list](https://frontshelf.io/docs/capabilities/gsc_notes_list): A site's Search Console notes - [gsc_notes_save](https://frontshelf.io/docs/capabilities/gsc_notes_save): Save a Search Console note - [gsc_notes_test](https://frontshelf.io/docs/capabilities/gsc_notes_test): The before/after test of Search Console notes - [gsc_overview](https://frontshelf.io/docs/capabilities/gsc_overview): Search Console across every site of the workspace - [gsc_pages_get](https://frontshelf.io/docs/capabilities/gsc_pages_get): Why one page's Search Console clicks moved - [gsc_pdfs_get](https://frontshelf.io/docs/capabilities/gsc_pdfs_get): A client PDF's state and download link - [gsc_pdfs_request](https://frontshelf.io/docs/capabilities/gsc_pdfs_request): Ask for a client PDF of a site's Search figures - [gsc_proposals_accept](https://frontshelf.io/docs/capabilities/gsc_proposals_accept): Accept one of the AI's suggestions - [gsc_proposals_accept_all](https://frontshelf.io/docs/capabilities/gsc_proposals_accept_all): Accept every open suggestion of one AI run - [gsc_proposals_dismiss](https://frontshelf.io/docs/capabilities/gsc_proposals_dismiss): Dismiss one of the AI's suggestions - [gsc_proposals_list](https://frontshelf.io/docs/capabilities/gsc_proposals_list): The AI's suggested page groups, keyword clusters and brand terms - [gsc_proposals_request](https://frontshelf.io/docs/capabilities/gsc_proposals_request): Ask the AI for page groups, keyword clusters and brand terms - [gsc_reports_export](https://frontshelf.io/docs/capabilities/gsc_reports_export): A whole Search Console report as a CSV link - [gsc_reports_get](https://frontshelf.io/docs/capabilities/gsc_reports_get): One page of a Search Console report - [gsc_schedules_list](https://frontshelf.io/docs/capabilities/gsc_schedules_list): A site's client report schedules - [gsc_schedules_save](https://frontshelf.io/docs/capabilities/gsc_schedules_save): Create or replace a site's weekly client email or monthly PDF schedule - [gsc_segments_delete](https://frontshelf.io/docs/capabilities/gsc_segments_delete): Delete a page group or keyword cluster - [gsc_segments_list](https://frontshelf.io/docs/capabilities/gsc_segments_list): A site's page groups and keyword clusters - [gsc_segments_preview](https://frontshelf.io/docs/capabilities/gsc_segments_preview): What a page group's or keyword cluster's conditions would match - [gsc_segments_save](https://frontshelf.io/docs/capabilities/gsc_segments_save): Save a page group or keyword cluster - [gsc_series_get](https://frontshelf.io/docs/capabilities/gsc_series_get): A site's Search Console figures over time - [gsc_sources_choose](https://frontshelf.io/docs/capabilities/gsc_sources_choose): Choose a site's primary Search Console property - [gsc_sources_list](https://frontshelf.io/docs/capabilities/gsc_sources_list): The Search Console properties a site reads from - [gsc_sources_refresh](https://frontshelf.io/docs/capabilities/gsc_sources_refresh): List the workspace's Search Console properties again - [gsc_table_export](https://frontshelf.io/docs/capabilities/gsc_table_export): A whole Search Console table as a CSV link - [gsc_table_get](https://frontshelf.io/docs/capabilities/gsc_table_get): One page of a site's Search Console rows - [gsc_watches_add](https://frontshelf.io/docs/capabilities/gsc_watches_add): Pin a page or query to a site's weekly drop check - [gsc_watches_list](https://frontshelf.io/docs/capabilities/gsc_watches_list): A site's drop alert settings and pinned pages and queries - [gsc_watches_remove](https://frontshelf.io/docs/capabilities/gsc_watches_remove): Unpin a page or query from a site's weekly drop check - [ideas_dismiss](https://frontshelf.io/docs/capabilities/ideas_dismiss): Set a growth idea aside - [ideas_get](https://frontshelf.io/docs/capabilities/ideas_get): Read one growth idea - [ideas_list](https://frontshelf.io/docs/capabilities/ideas_list): List a site's growth ideas - [ideas_report_live](https://frontshelf.io/docs/capabilities/ideas_report_live): Say your change for a growth idea is live - [ideas_report_pr](https://frontshelf.io/docs/capabilities/ideas_report_pr): Report the pull request your change for an idea is in - [ideas_restore](https://frontshelf.io/docs/capabilities/ideas_restore): Put a dismissed growth idea back - [ideas_start](https://frontshelf.io/docs/capabilities/ideas_start): Take a growth idea on - [Capabilities](https://frontshelf.io/docs/capabilities): Every capability an agent can call, the same on the CLI, MCP and REST. - [issues_acknowledge](https://frontshelf.io/docs/capabilities/issues_acknowledge): Say you are looking at an issue - [issues_get](https://frontshelf.io/docs/capabilities/issues_get): One issue, its work order and its verification history - [issues_ignore](https://frontshelf.io/docs/capabilities/issues_ignore): Set an issue aside as a deliberate won't-fix - [issues_list](https://frontshelf.io/docs/capabilities/issues_list): The open issue queue, worst first - [issues_report_deployed](https://frontshelf.io/docs/capabilities/issues_report_deployed): Say your fix for an issue is live - [issues_report_pr](https://frontshelf.io/docs/capabilities/issues_report_pr): Say you opened a pull request with your fix - [issues_unignore](https://frontshelf.io/docs/capabilities/issues_unignore): Put an ignored issue back in the queue - [ownership_get](https://frontshelf.io/docs/capabilities/ownership_get): Read whether a site's ownership is verified, and how to prove it - [pages_get](https://frontshelf.io/docs/capabilities/pages_get): One page of a committed crawl: its record, inlinks and every issue judged on it - [pages_list](https://frontshelf.io/docs/capabilities/pages_list): List the pages of a site's committed crawl, with filters, or the pages on one page filter - [pages_search](https://frontshelf.io/docs/capabilities/pages_search): Pages ranked by meaning against a query, over a committed crawl's embeddings - [reports_export](https://frontshelf.io/docs/capabilities/reports_export): A signed download link for one report of a crawl, or its PDF - [reports_list](https://frontshelf.io/docs/capabilities/reports_list): List the report catalogue, with a site's row counts - [rules_get](https://frontshelf.io/docs/capabilities/rules_get): One rule in full: what it checks, why, how to fix it, how a fix is verified - [rules_list](https://frontshelf.io/docs/capabilities/rules_list): List the rule catalogue: every check Frontshelf judges by - [schedules_create](https://frontshelf.io/docs/capabilities/schedules_create): Add a crawl schedule to a site - [schedules_delete](https://frontshelf.io/docs/capabilities/schedules_delete): Delete a crawl schedule - [schedules_list](https://frontshelf.io/docs/capabilities/schedules_list): List a site's crawl schedules - [schedules_update](https://frontshelf.io/docs/capabilities/schedules_update): Change a crawl schedule - [sessions_checkin](https://frontshelf.io/docs/capabilities/sessions_checkin): Check in with the task you are working on - [sessions_finish](https://frontshelf.io/docs/capabilities/sessions_finish): Tell your human you are done - [sessions_step](https://frontshelf.io/docs/capabilities/sessions_step): Tell your human the step you just took - [sitemaps_export](https://frontshelf.io/docs/capabilities/sitemaps_export): The site's XML sitemap from a crawl, as signed download links - [sites_create](https://frontshelf.io/docs/capabilities/sites_create): Add a site to audit; your human completes its checkout - [sites_get](https://frontshelf.io/docs/capabilities/sites_get): One site's health, latest crawl, trend and top issues - [sites_list](https://frontshelf.io/docs/capabilities/sites_list): List the sites this key can see, with health - [sites_tag](https://frontshelf.io/docs/capabilities/sites_tag): Set exactly the tags one site carries - [sites_verify](https://frontshelf.io/docs/capabilities/sites_verify): Start the check that proves your human owns a site - [tags_delete](https://frontshelf.io/docs/capabilities/tags_delete): Delete a site tag - [tags_list](https://frontshelf.io/docs/capabilities/tags_list): List the workspace's site tags - [tags_save](https://frontshelf.io/docs/capabilities/tags_save): Create a site tag, or save the next version of one - [vault_add_link](https://frontshelf.io/docs/capabilities/vault_add_link): Add a web page to a site's vault by its URL - [vault_add_text](https://frontshelf.io/docs/capabilities/vault_add_text): Add text to a site's vault as evidence - [vault_delete](https://frontshelf.io/docs/capabilities/vault_delete): Delete one item from a site's vault - [vault_get](https://frontshelf.io/docs/capabilities/vault_get): One vault item: its status, size and what extraction kept - [vault_list](https://frontshelf.io/docs/capabilities/vault_list): Every item in a site's vault, with its status - [vault_read](https://frontshelf.io/docs/capabilities/vault_read): One vault passage with its neighbours - [vault_rules_get](https://frontshelf.io/docs/capabilities/vault_rules_get): A site's writer's instructions, with the version a change sends - [vault_rules_set](https://frontshelf.io/docs/capabilities/vault_rules_set): Replace a site's writer's instructions, against the version you read - [vault_search](https://frontshelf.io/docs/capabilities/vault_search): Search a site's vault by meaning - [verifications_get](https://frontshelf.io/docs/capabilities/verifications_get): Read a verification and its verdict, waiting for it - [verifications_start](https://frontshelf.io/docs/capabilities/verifications_start): Ask the verifier to check a fix on the live site - [views_export](https://frontshelf.io/docs/capabilities/views_export): Export one Pages view of a crawl as a signed CSV or JSON Lines link - [CLI](https://frontshelf.io/docs/cli): The frontshelf CLI: local commands that see a site as search and AI crawlers do, install for coding agents, the cloud commands under frontshelf cloud, output and exit codes. - [Compare site checks](https://frontshelf.io/docs/compare): Changes and its Compare tab: what the latest site check found against the one before, the four differences per filter, change detection, URL mapping, and the change email after a scheduled check. - [Spelling, grammar and embeddings](https://frontshelf.io/docs/content): LanguageTool spelling and grammar per page, and embeddings for near-duplicates, low relevance, content clusters and semantic search, with coverage lines and costs. - [Advanced check settings](https://frontshelf.io/docs/crawl-settings): Every section of a site's Advanced check settings: budget and speed, scope, sources and lists, robots.txt, requests, logins, content, samples, thresholds, search, named schedules. - [How a site check works](https://frontshelf.io/docs/crawling): The stages of a Frontshelf site check, what the coverage counters mean, when a page gets rendered, the resource and external-link budgets, and the page limits. - [Custom extraction](https://frontshelf.io/docs/custom-extraction): Pull any data out of your pages with XPath, a CSS selector or a regular expression: define extractors once, every site check runs them, one column each. - [Custom search](https://frontshelf.io/docs/custom-search): Count a text or a regex on every page (source, visible text, head, body, an XPath or the content area) and filter All columns by the pages containing it. - [All columns](https://frontshelf.io/docs/explorer): Every URL a site check read, as data: quick views, the tabs of a widely used desktop SEO crawler with their filters and counts, segments, the page record, and export from any view. - [Getting started](https://frontshelf.io/docs): What Frontshelf is, the check-fix-verify loop, where each part of the product is documented, and how your coding agent connects: MCP, the CLI or the REST API. - [Google integrations](https://frontshelf.io/docs/integrations): Connecting Search Console and Google Analytics, what each pull fetches, URL Inspection and its daily budget, PageSpeed field data, the problems needing Google data. - [Lists and search previews](https://frontshelf.io/docs/lists): Check exactly a list of URLs, or judge a sheet of titles and descriptions without fetching: sources, what each mode discovers, how the score is labelled, the doors. - [Checking behind a login](https://frontshelf.io/docs/login-session): Hand Frontshelf a signed-in browser session as cookies from a bookmarklet, a cookies.txt or a HAR, and how a site check presents it, tests it, and tells you when it expires. - [MCP server](https://frontshelf.io/docs/mcp): POST /mcp with an Agent key: one tool, frontshelf (search, info, schema, call, learn), tools mode for Cursor, the capabilities behind call, and the laws. - [Verifying you own a site](https://frontshelf.io/docs/ownership): The three ways to prove ownership (DNS TXT, meta tag, well-known file), what verification unlocks, the weekly recheck, what happens if a proof goes. - [Reading a problem](https://frontshelf.io/docs/reading-issues): How to read a problem (in the app and on its public catalogue page), the health score's law, and a problem's steps from found to fixed. - [Reports and the PDF](https://frontshelf.io/docs/reports): A site's Reports screen: the branded PDF, read-only links, the weekly email and monthly PDF, the cover, and the 67 spreadsheets as streaming CSV or JSON. - [Verification](https://frontshelf.io/docs/verification): What a re-read is, what a verdict means (fixed, still present, inconclusive), why some checks wait for the next site check, and the laws that keep an agent from grading its own work. - [The Map](https://frontshelf.io/docs/visualise): Pages' Map view: the site by section, with the problems in each and the pages off Google, and a drawing of the links among its most linked-to pages. ## Blog - [Your agent cannot grade its own homework](https://frontshelf.io/blog/verification-loop): Coding agents patch sites in minutes. The missing piece is proof: why Frontshelf's verifier, an independent read of production, is the product. - [Two databases and a commit marker: how Frontshelf stores crawls forever](https://frontshelf.io/blog/immutable-crawls-commit-marker): Postgres holds transactional truth, ClickHouse holds append-only crawl history, a commit marker decides what readers may see: immutable audits.