Crawlio Browser

unavailable

Execute a JavaScript expression in the page context and return the result. Useful for reading page state, checking values, or performing calculations.

unavailable

Fill multiple form fields at once. Each field is targeted by ref (from browser_snapshot). Supports textbox, checkbox, radio, and combobox types. Returns post-fill snapshot.

unavailable

unavailable

unavailable

unavailable

unavailable

unavailable

Capture accessibility snapshot of the current page. Returns a compact ARIA tree with ref strings (e.g., [ref=e1]) for interactive elements. Use refs with browser_click, browser_type, browser_hover to target elements without constructing CSS selectors. Refs are valid until the page navigates or significant DOM changes occur — call browser_snapshot again after navigation. Options: interactive=true for 60-80% token reduction (interactive elements only), compact=true to remove empty structural branches, maxDepth to limit tree depth, selector to scope to a CSS subtree.

unavailable

unavailable

unavailable

unavailable

Cancel a running background execute job by jobId — terminates its sandbox worker. No-op if the job already finished.

Cancel an in-progress pick_element overlay on the connected tab and remove its highlight. Safe to call when no picker is active.

unavailable

Fetch and parse the robots.txt file for the current page's domain. Returns disallowed paths, sitemap URLs, and crawl-delay. Uses an in-page fetch() to access the file (preserves same-origin cookies/auth). Requires debugger attached.

Fetch and parse the sitemap.xml file for the current page's domain. Returns URL count, sample URLs, and last modified date. Tries /sitemap.xml first, then checks robots.txt for sitemap references. Requires debugger attached.

unavailable

Remove all Crawlio SERP overlay widgets (badges, header bar, sidebar) from the current page. Safe to call even if no overlay is present.

unavailable

unavailable

Compare raw HTML (pre-JavaScript) vs rendered DOM (post-JavaScript) to detect JS-dependent SEO content. Identifies differences in title, meta description, headings, canonical, structured data, links, and content size. Returns a risk assessment for search engine visibility. Requires debugger attached and network capture active (best when page was loaded with capture running).

Compile a recording session into a replayable SKILL.md markdown file. Takes the session JSON from stop_recording output and produces executable skill markdown with smart.* and bridge.send() calls. Pure computation — no browser access required.

Pin a browser tab for subsequent commands and start CDP capture. Three modes: (1) provide a URL: with the optional tabs grant Crawlio reuses a match, otherwise it creates a fresh owned tab; (2) provide a tabId to adopt a specific existing tab (tabs grant required); (3) no args to discover and pin the active tab (tabs grant required). Pass background:true to avoid stealing the user's active tab/window focus.

unavailable

POST /finding — create curated finding with evidence. Body: {title, url?, evidence?, synthesis?}

unavailable

Fetch Chrome User Experience Report (CrUX) field data for a URL — real-world Core Web Vitals from millions of Chrome users. Returns p75 values for LCP, CLS, INP, TTFB with good/needs-improvement/poor assessment. Requires a Google CrUX API key (pass as api_key or set CRUX_API_KEY env var). Low-traffic URLs may have no data — falls back to origin-level metrics automatically.

unavailable

DELETE /projects/{id} — delete a saved project by ID

Detect all fonts on the page. Returns loaded FontFace objects with family/weight/style/status/source/provider, @font-face CSS rules, font usage on key elements (computed font-family, font-size, font-weight), and a summary with provider and format breakdowns.

unavailable

Detect the page's structural regions from the live DOM: semantic tags (section/article/aside/header/main/footer/nav/form), ARIA landmark roles (banner/navigation/main/contentinfo/complementary/search/form/region), data-component/data-testid/data-section/data-block/data-module attributes, and PascalCase/BEM class-name hints. Returns a depth-capped SectionNode tree — each node has role, source, name (aria-label > data-component > first heading), a CSS selector with a selectorVerified uniqueness flag, bounding box, inViewport, interactiveCount, textLength, and headings. Sections deliberately emit…

Detect repeating data patterns (tables, lists, grids, card layouts) on the page. Native <table> elements are found first (strategy 'table-rows', with header rows promoted to column names by extract_table); div-soup layouts are found via class-frequency scoring with geometric row-height filtering (strategy 'sibling-repeated-blocks'). Returns ranked candidates with selectors, row counts, sample text, confidence (0..1), strategy, and warnings. Use extract_table to extract data from a candidate.

Detect technologies on the current page using CDP-captured signals matched against a curated fingerprint database (~200 technologies). Collects response headers, script URLs, JS globals, meta tags, cookies, and page URL — then runs additive confidence scoring (per-pattern weights summed, capped at 100). Returns technologies with numeric confidence, version, categories, and matched signals for full transparency. Requires debugger attached.

Compare the current ARIA snapshot against a previous one using Myers line-level diff. Returns a unified diff with addition/removal stats. If no baseline is provided, diffs against the last cached snapshot (from the most recent browser_snapshot call). Use this to verify interaction effects — click a button, then diff to see what changed.

unavailable

unavailable

unavailable

Navigate to a URL, capture browser intelligence (framework, network, console, DOM), and send enrichment to Crawlio — all in one call. Compound tool: connects tab if needed, navigates, waits for hydration, captures, and posts enrichment. Returns a compact summary (not the full capture).

Execute JavaScript code with access to the browser bridge, Crawlio HTTP client, and smart object. Use search() first to discover available commands and their parameters. IMPORTANT WARNINGS: - smart.screenshot() does NOT exist. For screenshots: bridge.send({ type: 'take_screenshot' }). - For structured page evidence, prefer smart.extractPage() — runs 7 ops in parallel with typed gaps[]. - capture_page returns a ~1KB shaped summary. For raw data, use stop_network_capture or get_console_logs. - Use smart.waitForIdle() instead of sleep(). Use smart.scrollCapture() instead of manual scroll loops. …

Export the active tab's cookies to ~/.crawlio/<domain>-session.json WITHOUT redaction, so you can reuse your own authenticated session across runs. Writes the raw CDP cookie array to disk — the file contains live credentials, so treat it like a password and keep it out of version control. The output path is confined to your home directory. Optionally filter by a specific domain.

POST /export — export downloaded site. Body: {format: folder|zip|singleHTML|warc, destinationPath}

Compound extraction: detect tables + extract each + collect JSON-LD structured data from the page. Returns all tables with columns/rows plus any schema.org/JSON-LD data. Combines detect_tables, extract_table, and JSON-LD extraction in one call.

Start a Crawlio crawl of the active tab's URL. Requires Crawlio to be running.

POST /extract — run Next.js RSC extraction pipeline. Body: {destinationPath?}

Extract structured data from a container element. Pass a selector from detect_tables. Native <table> containers get header-promoted column names (thead th) plus <name>_url companion columns for link-bearing cells; other containers use recursive path-keyed extraction with semantic column naming (link-first ordering, content hints like price/source_domain). Returns columns (name + original path + fill rate) and rows as key-value objects.

unavailable

unavailable

unavailable

Report browser-bridge command availability for the current tab, CDP domains, and optional permissions. Shows which commands are available, use a fallback, or are unavailable.

unavailable

Get current connection state: whether a tab is connected, CDP debugger attached, MCP server connected, and capture status. Use to check if connect_tab is needed.

unavailable

unavailable

GET /logs — recent log entries. Params: ?category=engine|download|parser|localizer|network|ui&level=debug|info|default|error|fault&limit=N

Get the current Crawlio crawl status — engine state, progress counters, sequence number.

Get URLs crawled by Crawlio with their status, content type, and size. Supports filtering by status (queued/downloading/completed/failed) and type (html/css/js/image). Returns paginated results.

GET /crawled-urls — downloaded URLs with status and pagination. Params: ?status=&type=&limit=&offset=

unavailable

unavailable

unavailable

unavailable

GET /downloads — all download items with status, HTTP code, bytes, content type

Get browser enrichment data stored in Crawlio for a URL (framework, network, console, DOM).

GET /enrichment — browser enrichment data (framework, network, console, DOM). Params: ?url=URL (optional)

GET /logs?level=error — error and fault-level log entries

GET /export/status — current export state and progress

GET /extract/status — current extraction state and progress

GET /failed-urls — failed download items with error messages

GET /findings — list curated findings. Params: ?host=&limit=

unavailable

Get the current idle debugger release setting (idleMs, enabled, minimumMs, defaultMs).

Poll a background execute job by jobId (returned by execute({ background: true })). Returns { status: running|done|error|cancelled, value?, console?, error?, ageMs, runtimeMs? }, or { status: 'not_found' } if unknown/expired (finished jobs are kept ~10 min). phase/percent come from reportPhase(name, percent) calls inside the job; phases is the recent timeline while it is still running.

GET /observations — append-only observation timeline. Params: ?host=&op=&source=&since=&limit=

unavailable

GET /projects/{id} — full project details by ID

Check the current recording status. Returns whether a recording is active, session ID, duration, page count, interaction count, and current page URL.

unavailable

unavailable

Get current SEO intelligence settings (enabled, autoOverlay, autoBadge).

GET /settings — current pending download settings and crawl policy

GET /site-tree — downloaded files as directory tree

unavailable

unavailable

List all of the user's open tabs across every window, most-recently-accessed first, with window and tab-group membership. Broader than list_tabs (includes non-HTTP and ungrouped tabs). Requires the tabs permission.

unavailable

unavailable

unavailable

unavailable

unavailable

Inject SEO overlay widgets into a Google SERP page using CDP + shadow DOM isolation. Requires debugger attached to a tab showing a Google search results page. Widgets: badge (per-result framework/perf), header (top bar with query info), sidebar (aggregated metrics panel). All rendering uses shadow DOM — no style conflicts with Google's page.

Inspect tracker runtime state via CDP Runtime.evaluate. Probes Facebook fbq (loaded, version, pixelIds, queue), GA4 dataLayer (events, gtag/ga presence), GTM containers (GTM-/G- IDs), and TikTok ttq (loaded, queue). Returns DataLayerState with null for absent trackers. Requires debugger attached — no content script needed.

List background execute jobs (running + recently finished). Returns { count, jobs: [{ jobId, status, ageMs, runtimeMs? }] }.

List the Chrome profiles that have connected to this server, marking which one currently holds the browser connection. A profile appears once its extension has connected at least once — Chrome gives an extension no way to see other profiles, so a profile that has never connected is invisible here.

GET /projects — all saved crawl projects

unavailable

List all open HTTP/HTTPS browser tabs with their IDs, URLs, titles, and which tab is currently connected. Use this to discover tab IDs for connect_tab.

POST /projects/{id}/load — load a saved project by ID

available

Manage extension-resident observation. Actions: training_start, training_status, training_stop, training_clear, training_artifacts; recording_start, recording_status, recording_stop, recording_clear, recording_artifacts (RecordingBundle aliases); monitor_start, monitor_status, monitor_results, monitor_stop, monitor_clear. Training and monitors continue while MCP is disconnected; stop materializes canonical files and confirmed clear deletes only Chrome-retained records.

Extract text from the current page using macOS Vision.framework OCR. Takes a screenshot via CDP, runs native text recognition (VNRecognizeTextRequest), and returns recognized text with confidence scores and bounding regions. Works on canvas elements, images, anti-scraping sites, and any visual content invisible to DOM extraction.

Parse captured network requests to detect tracking pixel fires from Facebook, GA4, TikTok, LinkedIn, and Pinterest. Extracts pixel IDs, event names, parameters, and vendor classification from NetworkEntry data. Operates on already-captured network data — no additional page interaction required.

POST /pause — pause the current crawl

Interactive element picker (extension-only — needs a human at the tab). Injects a hover-highlight overlay into the connected page; the user clicks an element and the tool returns a VERIFIED selector plus the 5-rail SelectorBundle (xpath/attribute/classChain/textContent/rolePlusText) forged by the @crawlio/selectors kernel and gated by resolvesExactlyTo. Set multiple=true to pick a repeating set (the user presses Enter to finish) and receive one generalized, verified array selector. Press Escape to cancel.

unavailable

unavailable

unavailable

List files in a canonical RecordingBundle directory without loading large bodies into the MCP response.

One-shot background canonical capture for automated flows: opens an owned tab without taking focus, starts a bundle, optionally waits for network idle, then stops and writes RecordingBundle v1 artifacts. For human demos prefer recording_start then recording_stop.

available

available

available

available

Validate that a canonical RecordingBundle directory contains the required v1 artifacts.

POST /recrawl — re-crawl specific URLs. Body: {urls: [string]}

unavailable

unavailable

Re-fire a previously captured network request with optional modifications. Requires active network capture. Specify the request by URL. Optionally override headers, body, or method. Returns the new response status, headers, and body. Useful for API testing, auth token replay, and form submission testing.

POST /resume — resume a paused crawl

List files in a robot-training artifact directory. Use after robot_training_stop to discover recording, network, body, state, and flows.jsonl files for robot training or OpenAPI synthesis.

available

available

available

available

POST /projects — save current crawl project. Body: {name?}

Search available commands by keyword — both browser automation (via bridge.send) and Crawlio HTTP endpoints (via crawlio.api). Returns matching command names, descriptions, and parameter schemas. Use this to discover what commands are available before writing execute() code.

Classify text against caller-provided labels using Crawlio's local semantic grounding seam. Provider outputs are advisory and provenance-stamped.

Find page candidate refs for a natural-language query using Crawlio's local semantic grounding seam. Provider outputs are advisory and include provenance; they do not authorize actions.

Run a comprehensive on-page SEO audit on the connected page via CDP. Extracts 11 dimensions: meta tags, heading structure, link profile, images, structured data, Open Graph, Twitter Cards, canonical/hreflang, robots directives, content metrics, and technical signals. Returns a scored audit with actionable issues. Requires debugger attached.

unavailable

unavailable

unavailable

unavailable

unavailable

Configure idle debugger release. After idleMs of no CDP activity the extension detaches the debugger so Chrome's "being debugged" banner disappears; the next tool call re-attaches automatically and restores the session (stealth, framework hooks, network capture). Pass idleMs: 0 to disable. Values below the 60000ms minimum are raised to it. Never fires during a recording, an active agent session, a pending dialog, installed request-interception rules, or active coverage.

unavailable

Configure passive SEO intelligence behavior. Controls auto-badge (SERP detection badge on Google SERPs), auto-overlay (automatic SERP overlay injection), and master enable/disable toggle. All settings are partial — omit a field to keep its current value.

unavailable

unavailable

unavailable

unavailable

unavailable

unavailable

POST /start — start a new crawl. Body: {url, urls?, destinationPath?}

unavailable

unavailable

unavailable

Start recording a browser session. Captures all tool interactions, page navigations, network requests, and console logs organized by page. Recording auto-stops after maxDurationSec (default 600s) or maxInteractions (default 500). A tab must be connected first.

POST /stop — stop the current crawl

unavailable

unavailable

unavailable

Stop the active recording session and return the complete session data. Returns a RecordingSession JSON with all pages, interactions (tool calls with args/results/timing), network requests, and console logs partitioned by page. If the recording was auto-stopped (duration/interaction limit/tab closed), returns that session.

unavailable

POST /enrichment/bundle — submit complete enrichment bundle. Body: {url, framework?, networkRequests?, consoleLogs?, domSnapshotJSON?}

POST /enrichment/console — submit console log entries. Body: {url, consoleLogs}

POST /enrichment/dom — submit DOM snapshot. Body: {url, domSnapshotJSON}

POST /enrichment/framework — submit framework detection data. Body: {url, framework}

POST /enrichment/network — submit captured network requests. Body: {url, networkRequests}

Move the browser connection to another Chrome profile. One profile is driven at a time; this chooses which. The current profile's extension is released and reconnects in the background, so switching back is immediate. Pass no profileId to clear the preference and accept whichever profile connects first.

unavailable

unavailable

unavailable

unavailable

unavailable

POST /capture — WebKit runtime capture (framework, network, console, DOM). Body: {url}

PATCH /settings — partial merge of settings and/or policy (idle-only). Body: {settings?, policy?}

Validate tracking pixel events against per-vendor parameter schemas. Checks for missing required/recommended parameters, invalid types, duplicate events, missing PageView, and possible event name typos. Uses parsed tracking data from parse_tracking_pixels. Returns issues with error/warning/info severity and actionable recommendations. Operates on already-captured network data.

Wait for all network requests to complete (idle detection). Uses CDP Network domain event tracking — catches ALL request types (fetch, XHR, images, CSS, fonts, scripts). Returns when no requests are pending for the specified idle window, or on timeout.