Crawlio Browser

Documentation | API Reference | Chrome Extension

MCP server that gives AI full control of a live Chrome browser via CDP. 150 tools with framework-aware intelligence, typed evidence infrastructure, tracking pixel analysis, technographic fingerprinting, SEO auditing, and confidence-tracked findings — captures what static crawlers can't see.

When to use Crawlio Browser

Use Crawlio Browser when your AI needs to interact with a real browser — SPAs, authenticated pages, dynamic content, JS-rendered frameworks. Unlike headless browser tools, Crawlio Browser connects to your actual Chrome via a lightweight extension, giving the AI access to your logged-in sessions, cookies, and full browser state.

Crawlio Browser vs headless browser tools: Headless tools launch a separate browser process. Crawlio Browser connects to your existing Chrome — no separate browser, no login flows, full access to your tabs and sessions.

[!WARNING] This is the trade-off, stated plainly: connecting your own Chrome is the feature, and it means an AI agent can act as you on every site you are logged into. It attaches Chrome's debugger to the tab you connect, so it can read cookies, storage, and page content for that session.

Review what you connect it to. Prefer a dedicated Chrome profile for agent work. Nothing is captured until you connect a tab, and you can see the exact tool surface before configuring anything by running npx crawlio-browser tools. Sites can opt out with <meta name="crawlio-agent" content="disable">, which the extension honors.

Quick Start

Install the Chrome Extension
Run the init wizard:
Terminal
```
npx crawlio-browser init
```

That's it. Auto-detects and configures 14 MCP clients: Claude Code, Cursor, VS Code, Codex, Gemini CLI, Claude Desktop, ChatGPT Desktop, Windsurf, Cline, Zed, Goose, OpenCode, MCPorter, and Cline CLI.

Init wizard options

Terminal

npx crawlio-browser init              # Default: code mode, stdio transport
npx crawlio-browser init --full       # Full mode (every tool exposed individually)
npx crawlio-browser init --portal     # Portal mode (persistent HTTP server)
npx crawlio-browser init --cloudflare # Add Cloudflare MCP (89 tools, no wrangler)
npx crawlio-browser init --dry-run    # Show what would happen
npx crawlio-browser init --yes        # Skip prompts (CI / scripted installs)
npx crawlio-browser init -a claude    # Target specific MCP client

As an Agent Plugin

The package is also an Agent Plugins v1 plugin, so a conformant client can load it directly instead of running the wizard — plugin.json at the root, the eleven skills under skills/, and the MCP server declared in mcp.json.

The product-facing crawlio-* workflows are folded into the eleven shipped skills. The similarly named definitions in agents/ remain repo-local development fixtures: they import src/evidence/* and loops/*, neither of which is part of the npm runtime. They are excluded from package.json instead of advertising a second, non-executable product surface.

Point the client at the installed package:

Code

node_modules/crawlio-browser        # after `npm install crawlio-browser`

mcp.json resolves the server through ${PLUGIN_ROOT}/dist/mcp-server/index.js, which is why the plugin has to be an installed package rather than an unpacked tarball — the server imports its dependencies at runtime, so a bare extract starts and then dies without answering.

Inspecting what it exposes

Terminal

npx crawlio-browser tools          # What code mode exposes (the default)
npx crawlio-browser tools --full   # Every tool, individually
npx crawlio-browser tools --json   # Machine-readable, for diffing across versions
npx crawlio-browser doctor         # Bridge, portal, native host, client configs
npx crawlio-browser --help         # All commands and options
npx crawlio-browser --version      # Version only

Both are read-only and run without a browser, an extension, or a network connection — you can see the whole surface before you configure any client. The numbers come from the same builders the server registers, so they cannot disagree with what your client receives.

Transport Modes

Mode	Command / URL	Protocol	Best For
stdio	`npx crawlio-browser`	JSON-RPC over stdin/stdout	Claude Desktop, Cursor, Windsurf — client manages process lifecycle
Portal (HTTP)	`POST http://127.0.0.1:3001/mcp`	MCP Streamable HTTP	Claude Code, ChatGPT Desktop — server survives session restarts
Portal (SSE)	`GET /sse` + `POST /message`	Server-Sent Events	Legacy clients needing SSE transport

Portal mode is recommended for Claude Code — the server persists across context compaction and session restarts. On macOS, --portal installs a launchd agent for auto-start on login.

Manual setup (any client)

Per-client manual config

Claude Desktop — add to claude_desktop_config.json:

config.json

{ "mcpServers": { "crawlio-browser": { "command": "npx", "args": ["-y", "crawlio-browser"] } } }

Claude Code (Portal Mode) — start npx crawlio-browser --portal, then add to .mcp.json:

config.json

{ "mcpServers": { "crawlio-browser": { "type": "http", "url": "http://127.0.0.1:3001/mcp" } } }

Claude Code (stdio):

Terminal

claude mcp add crawlio-browser -- npx -y crawlio-browser

Cursor — add to .cursor/mcp.json:

config.json

{ "mcpServers": { "crawlio-browser": { "command": "npx", "args": ["-y", "crawlio-browser"] } } }

Windsurf — add to Windsurf Settings > MCP:

config.json

{ "mcpServers": { "crawlio-browser": { "command": "npx", "args": ["-y", "crawlio-browser"] } } }

Cline (VS Code) — add to settings.json:

config.json

{ "cline.mcpServers": { "crawlio-browser": { "command": "npx", "args": ["-y", "crawlio-browser"] } } }

ChatGPT Desktop — Settings > Integrations > MCP: URL: http://127.0.0.1:3001/mcp | Type: Streamable HTTP

How It Works

Code

AI Client (stdio/http)  -->  MCP Server (Node.js)  -->  Chrome Extension (MV3)
                             crawlio-browser               WebSocket -> CDP

The MCP server communicates with the Chrome extension via WebSocket. The extension controls the browser through Chrome DevTools Protocol (CDP).

Capabilities

Framework-Aware Intelligence

Every execute call probes the browser for framework signatures and injects a shape-shifting smart object with framework-native accessors. React state, Vue reactivity, Next.js routing, Shopify cart data — 17 framework namespaces across 4 tiers, detected at runtime and rebuilt on every navigation. The AI doesn't query a generic DOM; it queries the framework's own data structures.

Evidence-Based Analysis

Method Mode adds higher-order methods and a typed evidence system on top of Code Mode. smart.extractPage() runs 7 parallel operations in a single call — page capture, performance metrics, security state, font detection, meta extraction, accessibility audit, and mobile-readiness check. Failed operations produce typed CoverageGap records instead of silent nulls. Findings created with smart.finding() get their confidence automatically adjusted when supporting data is missing. The result: structured, auditable research output with gap tracking and confidence propagation.

Session Recording & Replay

Record browser interactions as structured data, then compile them into reusable SKILL.md automations. 12 interaction tools are automatically intercepted during recording — clicks, typing, navigation, scrolling — each capturing args, result, timing, and page URL. One compileRecording() call converts the session into a deterministic automation script.

Robot Training

Capture human-guided browser demonstrations as replayable robot-training bundles. The default-mode observe lifecycle starts an event-driven recorder inside the extension; collection keeps running if the MCP process disconnects or restarts. On reconnection, training_stop exports the retained run and materializes the complete 13-file RecordingBundle for replay and API synthesis. Full mode keeps the existing robot_training_* names as compatibility aliases.

Page monitoring is resident for the same reason: an extension-owned background tab and Chrome alarm collect bounded ARIA snapshots while no MCP server is present. Training and monitor history share a 25 MiB local budget, with 20 completed training runs, 50 monitor jobs, 200 snapshots total, and 50 snapshots per monitor as count caps. Old completed data is evicted first; active work is never silently evicted. Work starts only through explicit MCP lifecycle actions; the same actions report status, stop collection, clear monitor snapshots, and—with an exact id plus explicit confirmation—delete a stopped training/recording record from Chrome while preserving its materialized files. The extension popup remains a connection and browser-access status surface. Storage values are keys-only unless the caller explicitly opts in. Monitor snapshots intentionally retain compact ARIA page text locally; starting a monitor should therefore be treated as consent to retain the visible content of that URL until it is cleared.

Auto-Settling & Actionability

Every mutative action (click, type, navigate, select_option) runs actionability checks before acting — polling visibility, dimensions, enabled state, and overlay detection. After the action, a progressive backoff settle delay ([0, 20, 100, 100, 500]ms) waits for DOM mutations to quiesce. The AI doesn't need manual sleep() calls between actions.

Several Tabs at Once

Any command that acts on a page takes an optional tabId from list_tabs. Omit it and the command runs on the connected tab, exactly as before; supply one and it runs on that tab instead, with the whole command surface available on each. Commands overlap, so two tabs can be driven at the same time:

server.ts

const [checkout, search] = await Promise.all([
  bridge.send({ type: "browser_snapshot", tabId: 42 }),
  bridge.send({ type: "browser_snapshot", tabId: 57 }),
]);

Targeting a tab never changes which tab connect_tab points at, so an agent working several tabs cannot reassign the one a human is watching. Frame selection, coverage sessions, and framework detection are per tab. Network capture is the exception — it records one tab at a time and says which tab holds it rather than interleaving two.

Chrome Profiles

An extension instance is confined to its own Chrome profile and cannot see any other, so with Crawlio enabled in more than one, commands land in whichever profile connected first. list_profiles shows the profiles that have connected and which is being driven; switch_profile moves the connection to another. One profile is driven at a time — the released extension reconnects in the background, so switching back is immediate.

Profiles identify themselves with a UUID minted into their own extension storage. It distinguishes a profile without describing it: no account, no email, no path, and no additional permission.

Selecting a profile keeps cooperating extensions out of each other's way — it is not a security boundary, since the id is asserted by the extension rather than proved. The bridge's existing protections are unchanged: one extension at a time, and each must prove the server holds the real bridge token before anything executes.

Architecture

A layered execution architecture where each layer absorbs a category of complexity that would otherwise fall on the model. The model sees four primary tools and a clean SDK. Everything beneath that surface is the runtime absorbing reality.

The layer worth understanding is the one that assembles itself. Detection runs against the live page on first use, and the smart object is built to match what that page turned out to be — smart.react.* exists only where React does. Nothing about the target is known at startup, so the surface is composed per tab rather than declared up front.

Code

                     ┌───────────────────────────────────┐
                     │        AI Model (LLM)             │
                     │  Writes code, reads errors, loops  │
                     └───────────────┬───────────────────┘
                                     │  search, execute, observe, connect_tab (+ 3 job tools)
                                     ▼
┌─────────────────────────────────────────────────────────────────┐
│                     Crawlio Browser runtime                      │
│                                                                  │
│  ┌────────────────────────────────────────────────────────────┐  │
│  │  METHOD MODE                                               │  │
│  │  Behavioral protocol + higher-order methods                │  │
│  │  scrollCapture · waitForIdle · extractPage · comparePages  │  │
│  │  detectTables · extractTable · waitForNetworkIdle ·        │  │
│  │  extractData                                               │  │
│  │                                                            │  │
│  │  ↳ Absorbs: behavioral variance, ad-hoc composition,      │  │
│  │    inconsistent output shapes, data extraction patterns    │  │
│  ├────────────────────────────────────────────────────────────┤  │
│  │  POLYMORPHIC CONTEXT                                       │  │
│  │  17 framework namespaces, injected Just-In-Time            │  │
│  │  react · vue · angular · nextjs · shopify · ...            │  │
│  │                                                            │  │
│  │  ↳ Absorbs: framework opacity, minified code,             │  │
│  │    devtools hook complexity                                │  │
│  ├────────────────────────────────────────────────────────────┤  │
│  │  ACTIONABILITY ENGINE                                      │  │
│  │  7 core smart methods with built-in resilience             │  │
│  │  click · type · navigate · waitFor · evaluate ·            │  │
│  │  snapshot · rebuild                                        │  │
│  │                                                            │  │
│  │  ↳ Absorbs: DOM timing, hydration delays, CSS animations, │  │
│  │    disabled states, overlapping elements                   │  │
│  ├────────────────────────────────────────────────────────────┤  │
│  │  TETHERED IPC BRIDGE                                       │  │
│  │  WebSocket ↔ Chrome extension, message queue,              │  │
│  │  heartbeat, auto-reconnect, stale detection                │  │
│  │                                                            │  │
│  │  ↳ Absorbs: connection drops, tab refreshes,              │  │
│  │    port conflicts, extension lifecycle                     │  │
│  ├────────────────────────────────────────────────────────────┤  │
│  │  COMMAND CHANNEL                                           │  │
│  │  bridge.send → CDP browser control via the extension       │  │
│  │  crawlio.*   → Crawlio HTTP endpoints                      │  │
│  │  Live searchable catalog: browser + Crawlio HTTP           │  │
│  └────────────────────────────────────────────────────────────┘  │
└─────────────────────────────────────────────────────────────────┘
                                     │
                                     ▼
                     ┌───────────────────────────────────┐
                     │         Live Chrome Browser        │
                     │   Persistent session, real DOM,    │
                     │   framework runtime, user state    │
                     └───────────────────────────────────┘

What Each Layer Absorbs

Layer	Without It	With It
Tethered IPC Bridge	Script crashes on tab refresh, pending commands lost on reconnect, port conflicts on startup	Resilient WebSocket with message queue (100-msg capacity), heartbeat stale detection (15s intervals), auto-reconnect with drain
Actionability Engine	`click('#btn')` fires before the button renders, during CSS transitions, or while an overlay covers it	Progressive polling (exists → has dimensions → visible → not disabled → not obscured) with `[0, 20, 100, 100, 500]ms` backoff
Polymorphic Context	Model sees minified `<div>` elements; reading React state requires knowing exact hook paths, renderer maps, and fiber root API	Runtime probes live JS environment, recognizes 64 frameworks and attaches up to 17 matching namespaces (`smart.react.getVersion()`, `smart.nextjs.getData()`)
Method Mode	Model composes primitives ad-hoc — inconsistent scroll loops, missed edge cases, varying return shapes	18 tested methods encode correct patterns; behavioral protocol constrains workflow

Execution Lifecycle

Discovery — Model calls search("page capture performance") and gets documentation for relevant commands
Framework Detection — Runtime probes the live DOM, detects active frameworks, constructs polymorphic smart object with appropriate namespaces
Scope Assembly — Model's code is compiled into an async function with injected parameters: bridge (the browser command channel), crawlio (HTTP client), sleep, TIMEOUTS, smart (7 core + 18 higher-order methods + up to 17 framework namespaces), compileRecording
Execution — Method Mode methods compose the lower layers: extractPage() fires 7 parallel bridge.send() calls; click() runs the actionability engine; react.getVersion() evaluates framework-specific expressions
Error Recovery (Agentic REPL) — On failure, the browser stays in the exact state that produced the error. The model reads the structured error, adjusts, and calls execute again. Framework cache persists — no re-detection unless URL changed

Design Principles

Absorb complexity downward — Every category of difficulty (connection management, DOM timing, framework detection, multi-step composition) is handled by the layer best equipped for it. The model only encounters the clean interface at the top.
Shape the SDK to the target — The polymorphic context system detects what the page is and reshapes available methods to match. The model writes against a stable interface; the runtime adapts underneath.
Preserve state across cycles — The tethered architecture means the model can fail, learn, and retry against the same live environment — transforming error handling from "restart from scratch" into "adjust and continue."

How It Compares

Code Mode is Cloudflare's idea and a good one: present tools as a typed API and let the model write code against it, because models have seen far more code than tool calls. Crawlio Browser applies that pattern to a target it was not built for — a live browser holding your session. This is an MCP server, not an alternative to MCP.

Dimension	Standard MCP	Cloudflare Code Mode	Crawlio Browser
Tools in context	50-100+ schemas	2 (`search`, `execute`)	7 (4 primary + 3 async job tools)
Execution environment	N/A (tool calls)	V8 isolate (stateless)	Local async sandbox, tethered to a live browser
DOM access	Via individual tool calls	None	Live, persistent, framework-aware
Framework awareness	None	None	up to 17 namespaces, attached per page
Action resilience	Model must handle timing	N/A (no DOM)	Built-in actionability polling + settle delays
Error recovery	Re-call individual tool	Re-create isolate	Re-run against the state that produced the error
Multi-step patterns	Model improvises	Model writes loops	18 tested higher-order methods + behavioral protocol

The row that matters is error recovery. An isolate is disposable by design — that is what makes it safe to run untrusted code, and it means a failure discards the state that caused it. Here execution is tethered to a real tab, so when code fails the browser is still sitting in exactly the situation that broke it: same scroll position, same modal, same half-filled form. The model reads the error and runs again against that. Neither approach is better in general; they are answers to different problems, and the sandboxing guarantees Cloudflare gets from V8 isolates are genuinely stronger than what a local async sandbox provides.

Read the full architecture guide →

Two Modes

Code Mode (4 primary tools) — default

Collapses 150 tools into four high-level tools, plus get_job_result, list_jobs, and cancel_job for async execution — so tools/list reports

That makes the tools/list payload 85% smaller, measured by serializing the result each mode actually returns rather than estimated from a tool count. Code mode is not free: execute and search carry long descriptions, so the measured reduction is smaller than a naive tools-only ratio suggests. Check it yourself — npx crawlio-browser tools --json prints both surfaces without connecting to anything.

Tool	Description
`search`	Discover available commands by keyword
`execute`	Run async JS with `bridge`, `crawlio`, `smart`, `sleep`, and `compileRecording` in scope
`observe`	Start/query/stop extension-resident training, recording, and page monitors
`connect_tab`	Connect to a browser tab

server.ts

// Navigate and screenshot
await bridge.send({ type: 'browser_navigate', url: 'https://example.com' }, 30000);
await sleep(2000);
const screenshot = await bridge.send({ type: 'take_screenshot' }, 10000);
return screenshot;

Full Mode (150 tools)

Every tool exposed directly to the LLM. Enable with --full:

Terminal

npx crawlio-browser init --full

Smart Object

In Code Mode, the smart object provides framework-aware helpers with auto-waiting and actionability checks.

Core Methods

Method	Description
`smart.evaluate(expression)`	Execute JS in the page via CDP
`smart.click(selector, opts?)`	Auto-waiting click with 500ms settle
`smart.type(selector, text, opts?)`	Auto-waiting type with 300ms settle
`smart.navigate(url, opts?)`	Navigate with 1000ms settle
`smart.waitFor(selector, timeout?)`	Poll until element is actionable
`smart.snapshot()`	Accessibility tree snapshot
`smart.rebuild()`	Re-detect frameworks and reattach namespaces for the current page

Higher-Order Methods

Method	Description
`smart.scrollCapture(opts?)`	Scroll to bottom, capturing screenshots at each position. Handles stuck-scroll detection, bottom detection, section capping, and scroll reset.
`smart.waitForIdle(timeout?)`	MutationObserver-based idle detection — waits for 500ms quiet window. Timeout hard-capped at 15s. Replaces blind `sleep()` calls.
`smart.extractPage(opts?)`	7 parallel operations in one call — page capture, performance, security, fonts, meta, accessibility, mobile-readiness. Returns typed `PageEvidence` with `CoverageGap[]` for anything that failed.
`smart.comparePages(urlA, urlB)`	Navigates to both URLs, runs `extractPage()` on each, returns a `ComparisonScaffold` with 11 dimensions, shared/missing fields, and comparable metrics.

Typed Evidence

Methods for structured analysis findings with confidence propagation:

Method	Description
`smart.finding(data)`	Create a validated `Finding` with claim, evidence, sourceUrl, confidence, and method. Rejects malformed input with specific errors.
`smart.findings()`	Get all session-accumulated findings (returns a copy)
`smart.clearFindings()`	Reset session findings and coverage gaps

When a finding's dimension matches an active coverage gap, confidence is automatically capped:

Input Confidence	Active Gap	Output
`high`	`reducesConfidence: true`	`medium` + `confidenceCapped: true`
`medium`	`reducesConfidence: true`	`low` + `confidenceCapped: true`
`low`	any	`low` (floor)
any	no matching gap	unchanged

Framework Namespaces

When a framework is detected, the smart object exposes framework-specific helpers:

React — smart.react

Method	Returns
`getVersion()`	Version string and bundle type
`getRootCount()`	Number of React root components
`hasProfiler()`	Whether profiler is available
`isHookInstalled()`	Whether DevTools hook is installed

Vue.js — smart.vue

Method	Returns
`getVersion()`	Vue version string
`getAppCount()`	Number of Vue app instances
`getConfig()`	App config object
`isDevMode()`	Whether DevTools is enabled

Angular — smart.angular

Method	Returns
`getVersion()`	ng-version attribute value
`isDebugMode()`	Whether debug APIs available
`isIvy()`	Whether Ivy compiler is active
`getRootCount()`	Number of Angular root elements
`getState()`	Full state object

Svelte — smart.svelte

Method	Returns
`getVersion()`	Svelte version string
`getMeta()`	Svelte metadata object
`isDetected()`	Whether Svelte is detected

Redux — smart.redux

Method	Returns
`isInstalled()`	Whether Redux DevTools is installed
`getStoreState()`	Full store state

Alpine.js — smart.alpine

Method	Returns
`getVersion()`	Alpine version string
`getStoreKeys()`	Store object keys
`getComponentCount()`	Count of `[x-data]` components

Next.js — smart.nextjs

Method	Returns
`getData()`	`__NEXT_DATA__` object
`getRouter()`	Router state (pathname, query, asPath)
`getSSRMode()`	SSR mode (hybrid, app-router, static)
`getRouteManifest()`	Current page data

Nuxt — smart.nuxt

Method	Returns
`getData()`	`__NUXT__` object
`getConfig()`	App config
`isSSR()`	Whether server-rendered

Remix — smart.remix

Method	Returns
`getContext()`	`__remixContext` object
`getRouteData()`	Loader data from state

Shopify — smart.shopify

Method	Returns
`getShop()`	Shop metadata (theme, locale, currency)
`getCart()`	Shopping cart object

WordPress — smart.wordpress

Method	Returns
`isWP()`	Whether WordPress is present
`getRestUrl()`	REST API endpoint
`getPlugins()`	List of active plugins

More frameworks — Gatsby, WooCommerce, Laravel, Django, Drupal, jQuery

Namespace	Methods
`smart.gatsby`	`getData()`, `getPageData()`
`smart.woocommerce`	`getParams()`
`smart.laravel`	`getCSRF()`
`smart.django`	`getCSRF()`
`smart.drupal`	`getSettings()`
`smart.jquery`	`getVersion()`

Method Mode

Code Mode asks the model to write code. Method Mode gives that code a tested vocabulary:

18 higher-order methods that encode the multi-step patterns a model

would otherwise improvise.

javascript

await smart.extractTable(selector)   // not a hand-rolled scrape loop
await smart.scrollCapture()          // not a guessed scroll cadence

It is a domain layer over Code Mode, not a replacement: the tool surface does not change, the model still sees the same four primary tools, and the same 182-command catalog sits underneath. What changes is what happens inside execute.

The surface is assembled per page. Detection recognizes 64 frameworks and attaches up to 17 matching namespaces, so smart.react.* exists only where React does — the model never has to ask what the page is built with, or guess at hook paths and fiber internals to find out.

And because execution is tethered to a live tab rather than a disposable isolate, a failure leaves the browser in the exact state that produced it. The model reads the structured error and runs again against that state, rather than rebuilding the situation from scratch.

The Maturity Ladder

Layer	Optimizes For	Behavioral Variance	Evidence Quality
Raw MCP (150 tools)	Completeness	High — flat tool list, no composition guidance	None — unstructured text
Code Mode (7 tools)	Token efficiency	Medium — right primitives, ad-hoc composition	None — model-defined shapes
Method Mode (+ 18 methods + protocol)	Consistency	Low — proper methods, protocol constraints	Convention — `{ finding, evidence, url }`
+ typed evidence (gaps + confidence)	Correctness	Minimal — typed schemas, tool-enforced findings	Structural — typed records, gap tracking, confidence propagation

Architecture

Code

┌────────────────────────────────────────────────────────────┐
│                      execute sandbox                       │
│                                                            │
│  ┌──────────────────────────────────────────────────────┐  │
│  │  Behavioral Protocol  (web-research skill)           │  │
│  │  Acquire → Normalize → Analyze                       │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │  Evidence Infrastructure                             │  │
│  │  finding() · findings() · clearFindings()            │  │
│  │  Typed records · Coverage gaps · Confidence prop.    │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │  Higher-Order Methods  [18]                          │  │
│  │  scrollCapture · waitForIdle · extractPage ·         │  │
│  │  comparePages · detectTables · extractTable ·        │  │
│  │  waitForNetworkIdle · extractData · detectSections · │  │
│  │  detectTechnologies · parseTrackingPixels · ...      │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │  Smart Core  [7 methods]                             │  │
│  │  evaluate · click · type · navigate · waitFor ·      │  │
│  │  snapshot · rebuild                                  │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │  Framework Namespaces  [up to 17, attached per page] │  │
│  │  react · vue · angular · nextjs · shopify · ...      │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │  bridge.send()  — the browser command channel        │  │
│  └──────────────────────────────────────────────────────┘  │
└────────────────────────────────────────────────────────────┘

Each layer up encodes more domain knowledge. bridge.send({ type: "capture_page" }) captures a page. smart.extractPage() captures a page AND runs performance metrics, security state, font detection, accessibility analysis, and mobile-readiness checks in parallel — seven operations, one call, graceful failure on supplementary data, typed gaps for anything that fails.

Evidence Infrastructure

Coverage Gaps — When supplementary operations in extractPage() fail, they don't silently return null. A typed gap is recorded with the dimension, reason, impact, and whether it reduces confidence on related findings:

javascript

// Example gap from a failed performance metrics call
{ dimension: "performance", reason: "CDP domain disabled", impact: "method-failed", reducesConfidence: true }

Tool-Enforced Findings — smart.finding() validates every field at the tool level. The model cannot produce a finding without meeting the schema — it either returns a valid Finding or gets a clear error. Findings accumulate across execute calls within a session via smart.findings().

Session Aggregation — Findings and coverage gaps persist across execute calls. A model can make findings across multiple calls, then retrieve the full set with smart.findings(). Reset with smart.clearFindings().

End-to-End Example: Competitive Audit

server.ts

// 1. Extract and compare both sites (scaffold + gaps included)
const comparison = await smart.comparePages(
  'https://acme.com',
  'https://rival.com'
);

// 2. Make findings — confidence auto-adjusts based on data availability
smart.finding({
  claim: 'Rival loads 2.3x faster on Largest Contentful Paint',
  evidence: [
    `Acme LCP: ${comparison.siteA.performance?.webVitals?.lcp}ms`,
    `Rival LCP: ${comparison.siteB.performance?.webVitals?.lcp}ms`,
  ],
  sourceUrl: 'https://acme.com',
  confidence: 'high',
  method: 'comparePages + extractPage performance metrics',
  dimension: 'performance',  // if perf data failed, confidence caps to "medium"
});

smart.finding({
  claim: 'Acme has 12 images without alt text; Rival has 0',
  evidence: [
    `Acme imagesWithoutAlt: ${comparison.siteA.accessibility?.imagesWithoutAlt}`,
    `Rival imagesWithoutAlt: ${comparison.siteB.accessibility?.imagesWithoutAlt}`,
  ],
  sourceUrl: 'https://acme.com',
  confidence: 'high',
  method: 'comparePages + extractPage accessibility summary',
  dimension: 'accessibility',
});

// 3. Capture visual evidence
await smart.navigate('https://acme.com');
await smart.waitForIdle();
const acmeVisuals = await smart.scrollCapture({ maxSections: 5 });

// 4. Return accumulated session findings + visual evidence
return {
  findings: smart.findings(),
  scaffold: comparison.scaffold,
  gaps: { acme: comparison.siteA.gaps, rival: comparison.siteB.gaps },
  visualEvidence: { acme: acmeVisuals.sectionCount + ' sections captured' },
};

Examples

Navigate, extract, and analyze

server.ts

// Connect to active tab, extract structured page evidence
const page = await smart.extractPage();
const finding = smart.finding({
  claim: `Site uses ${page.capture.framework?.name || 'no detected framework'}`,
  evidence: [`Framework: ${JSON.stringify(page.capture.framework)}`],
  sourceUrl: page.meta?.canonical || 'active tab',
  confidence: 'high',
  method: 'extractPage framework detection',
});
return { page: page.meta, finding };

Mobile emulation + screenshot

server.ts

// Emulate iPhone and capture
await bridge.send({ type: 'emulate_device', device: 'iPhone 14' }, 10000);
await smart.navigate('https://example.com');
await smart.waitForIdle();
const screenshot = await bridge.send({ type: 'take_screenshot' }, 10000);
return screenshot;

Record and compile automation

server.ts

// Record a browser session, then compile to reusable skill
await bridge.send({ type: 'start_recording' }, 10000);
await smart.navigate('https://example.com');
await smart.click('button.submit');
await smart.type('#email', 'test@example.com');
const session = await bridge.send({ type: 'stop_recording' }, 10000);
return compileRecording(session.session, 'signup-flow');

Intercept and mock network

javascript

// Block analytics, mock API response
await bridge.send({
  type: 'browser_intercept',
  pattern: '*analytics*',
  action: 'block'
}, 10000);
await bridge.send({
  type: 'browser_intercept',
  pattern: '*/api/user',
  action: 'mock',
  body: JSON.stringify({ name: 'Test User' }),
  statusCode: 200
}, 10000);
await smart.navigate('https://example.com');
return await smart.snapshot();

Session Recording

Record browser sessions as structured data, then compile them into reusable automation skills. 12 interaction tools are automatically intercepted during recording (click, type, navigate, scroll, etc.), capturing args, result, timing, and page URL.

server.ts

// In code mode: record, interact, compile
await bridge.send({ type: 'start_recording' }, 10000);
// ... interact with the page ...
const session = await bridge.send({ type: 'stop_recording' }, 10000);
const skill = compileRecording(session.session, 'my-automation');
return skill;

In full mode, recording is available as 4 individual tools: start_recording, stop_recording, get_recording_status, and compile_recording.

Auto-Settling

Mutative tools (browser_click, browser_type, browser_navigate, browser_select_option) use actionability checks:

Pre-flight: Polls element visibility, stability, and enabled state before acting
Action: Dispatches the CDP command
Post-settle: Waits for DOM mutations to quiesce with progressive backoff [0, 20, 100, 100, 500]ms

This means the AI doesn't need to manually add sleep() or waitFor() calls between actions — the tools handle SPA rendering delays automatically.

Framework Detection

Detects 64 technologies across 4 tiers using globals, DOM markers, meta tags, HTTP headers, and script URLs:

Tier	Frameworks	Signal Strength
Meta-frameworks	Next.js, Nuxt, SvelteKit, Remix, Gatsby	Unique globals + parent detection
Core	React, Vue.js, Angular, Svelte, Astro, Qwik, SolidJS, Lit, Preact	Globals + DOM markers
CMS & Platforms	WordPress, Shopify, Webflow, Squarespace, Wix, Drupal, Magento, Ghost, Bubble	Meta tags + globals
Libraries & Tools	jQuery, Bootstrap, Tailwind CSS, Alpine.js, HTMX, Turbo, Stencil, Redux, Ember.js, Backbone.js	DOM + globals

Multi-framework detection returns a primary framework (meta-framework takes priority) plus a subFrameworks array for the full stack.

Tools Reference

The table below is a hand-written guide to the commonly used tools, not the complete set. For the full, current surface — all 150 of them, read from the server itself — run npx crawlio-browser tools --full.

Tool reference — Connection, Capture, Navigation, Network, Storage, Emulation, Tracking, SEO, and more

Connection & Status

Tool	Description
`connect_tab`	Connect to a browser tab by URL, tab ID, or active tab
`disconnect_tab`	Disconnect from the current tab
`list_tabs`	List all open tabs with IDs and URLs
`get_connection_status`	Check CDP connection state
`reconnect_tab`	Force reconnect to fix stale connections
`get_capabilities`	Report live browser-bridge availability by tab, CDP domain, and permission state

Page Capture

Tool	Description
`capture_page`	Full capture: framework + network + console + DOM
`detect_framework`	Detect JS framework and version
`start_network_capture`	Start recording network requests
`stop_network_capture`	Stop recording and return captured requests
`get_console_logs`	Get console logs (errors, warnings, info)
`get_cookies`	Get cookies (sensitive values redacted)
`get_dom_snapshot`	Simplified DOM tree with shadow DOM and iframe support
`take_screenshot`	Viewport/full-page/element image (JPEG default, PNG on request)
`get_response_body`	Get response body for a captured network request

Navigation & Interaction

Tool	Description
`browser_navigate`	Navigate to a URL (auto-settle)
`browser_click`	Click element by CSS selector (auto-settle, left/right/middle, modifiers)
`browser_double_click`	Double-click element
`browser_type`	Type text into element (auto-settle)
`browser_press_key`	Press keyboard key (Enter, Tab, Escape, shortcuts)
`browser_hover`	Hover over element
`browser_select_option`	Select `<option>` by value (auto-settle)
`browser_scroll`	Scroll page or element
`browser_drag`	Drag from one element to another
`browser_file_upload`	Upload files to `<input type="file">`
`browser_wait`	Wait N milliseconds
`browser_wait_for`	Wait for element state (visible, hidden, attached, detached)

Network

Tool	Description
`browser_intercept`	Block, modify headers, or mock responses for URL patterns
`emulate_network`	Throttle network (offline, 3G, 4G, WiFi presets)
`set_cache_disabled`	Disable/enable browser cache
`set_extra_headers`	Add custom headers to all requests
`get_websocket_connections`	List active WebSocket connections
`get_websocket_messages`	Get WebSocket message history

Frames & Tabs

Tool	Description
`get_frame_tree`	Get frame hierarchy (main + iframes)
`switch_to_frame`	Switch execution context to iframe
`switch_to_main_frame`	Switch back to main frame
`create_tab`	Create new tab with URL
`close_tab`	Close tab by ID
`switch_tab`	Focus a tab by ID

Cookies & Storage

Tool	Description
`set_cookie`	Set cookie (supports httpOnly via CDP)
`delete_cookies`	Delete cookies by name/domain/path
`get_storage`	Read localStorage or sessionStorage
`set_storage`	Write storage item
`clear_storage`	Clear all storage items
`get_databases`	List IndexedDB databases
`query_object_store`	Query IndexedDB object store
`clear_database`	Clear or delete IndexedDB database

Dialogs

Tool	Description
`get_dialog`	Get pending JS dialog (alert/confirm/prompt)
`handle_dialog`	Accept or dismiss dialog

Emulation

Tool	Description
`set_viewport`	Set viewport dimensions
`set_user_agent`	Override User-Agent string
`emulate_device`	Emulate device (iPhone, iPad, Pixel, Galaxy, Desktop)
`set_geolocation`	Override geolocation coordinates
`set_stealth_mode`	Anti-detection mode (opt-in, patches webdriver fingerprint)

Security

Tool	Description
`get_security_state`	TLS certificate details, protocol, cipher
`ignore_certificate_errors`	Ignore cert errors for staging environments

Service Workers

Tool	Description
`list_service_workers`	List all service worker registrations
`stop_service_worker`	Stop/unregister a service worker
`bypass_service_worker`	Bypass service workers for network requests

DOM Manipulation

Tool	Description
`set_outer_html`	Replace element's HTML
`set_attribute`	Set element attribute
`remove_attribute`	Remove element attribute
`remove_node`	Remove element from DOM

CSS & JS Coverage

Tool	Description
`start_css_coverage` / `stop_css_coverage`	Track which CSS rules are used
`start_js_coverage` / `stop_js_coverage`	Track which JS code is executed
`get_computed_style`	Get resolved CSS properties for element
`force_pseudo_state`	Force :hover, :focus, :active states

Performance & Memory

Tool	Description
`get_performance_metrics`	Chrome metrics + Web Vitals (LCP, CLS, FID)
`get_dom_counters`	Count DOM nodes, documents, event listeners
`force_gc`	Force garbage collection
`take_heap_snapshot`	V8 heap snapshot summary

PDF & Accessibility

Tool	Description
`print_to_pdf`	Generate PDF (custom paper, margins, orientation)
`get_accessibility_tree`	Accessibility tree for screen-reader audit

Targets & Contexts

Tool	Description
`get_targets`	List all Chrome targets (pages, workers, extensions)
`attach_to_target`	Attach CDP session to any target
`create_browser_context`	Create isolated (incognito-like) context

Visual Debug

Tool	Description
`highlight_element`	Highlight element with colored overlay
`show_layout_shifts`	Visualize CLS regions
`show_paint_rects`	Visualize paint/repaint areas

Session Recording

Tool	Description
`start_recording`	Begin recording browser session
`stop_recording`	Stop recording and return session data
`get_recording_status`	Check recording state
`compile_recording`	Compile session into SKILL.md automation

Robot Training

Tool	Description
`robot_training_start`	Start a fresh monitored demonstration run
`robot_training_status`	Query extension-retained runs and recording state
`robot_training_stop`	Stop/export a resident run and persist all bundle artifacts
`robot_training_clear`	Confirm deletion of one stopped extension-retained run; preserve artifact files
`robot_training_artifacts`	List files in a robot-training artifact directory
`monitor_page`	Start/query/stop extension-resident recurring page monitors

Crawlio App Integration

Optional — requires Crawlio.app running locally.

ControlServer authentication is automatic and local-only: the MCP reads CRAWLIO_MCP_TOKEN when explicitly set, otherwise Crawlio.app's mode-0600 ~/Library/Logs/Crawlio/mcp.token, and sends it only to the discovered 127.0.0.1 ControlServer. The value is never returned in MCP results or written to logs.

Tool	Description
`extract_site`	Start a Crawlio crawl of the active tab's URL
`get_crawl_status`	Get crawl progress and status
`get_enrichment`	Get browser enrichment data
`get_crawled_urls`	Get crawled URLs with status and pagination
`enrich_url`	Navigate + capture + submit enrichment in one call

Requirements

Node.js >= 18
Chrome (or Chromium) with the Crawlio for Chrome extension installed
Crawlio.app (optional) — for site crawling and enrichment

Permission floor

The production extension has no standing host access. Its required permissions are debugger (the CDP control plane), storage (bridge/session settings and durable resident metadata), and alarms (reconnect, idle-release, and resident monitor wakeups). The dedicated onboarding page asks once for every optional capability declared by the active build. In production that is tabs, nativeMessaging, and http://127.0.0.1/*, covering tab discovery/adoption, authenticated local token provisioning, and loopback bridge discovery.

Onboarding is the only surface that can open Chrome's permission prompt. The extension popup and MCP tools only report missing access and route the user back to onboarding. connect_tab({url}), agent-owned tabs, robot training, and resident monitoring can still create and control their own tabs if the tabs metadata grant is denied. Crawlio does not request <all_urls>, activeTab, tabGroups, or unlimitedStorage.

Build from Source

bash

git clone https://github.com/Crawlio-app/crawlio-browser.git
cd crawlio-browser
npm install
npm run build          # selectors → semantic-grounding → MCP server → extension

Build output lands in dist/mcp-server/ (ESM bundle) and dist/extension/ (the unpacked MV3 extension).

To develop against the extension, build the dev variant and load it unpacked:

Terminal

npm run build:dev      # → dist/extension-dev, with __DEV__ logging enabled

Then open chrome://extensions, enable Developer mode, choose Load unpacked, and select dist/extension-dev. Reload the extension there after each rebuild.

Terminal

npm test               # vitest
npm run typecheck      # server + extension TypeScript projects

Resources

Contributing

Contributions are welcome — see CONTRIBUTING.md for the build and development loop, and CODE_OF_CONDUCT.md for community expectations. Please report security vulnerabilities privately via SECURITY.md rather than in a public issue.

License

Apache-2.0 — see LICENSE.

The selector kernel in packages/selectors is MIT-licensed: it contains code ported from Selector Forge (MIT) and Chromium DevTools (BSD-3-Clause). See packages/selectors/LICENSE and THIRD_PARTY_NOTICES.md for full attributions.