Browser Infrastructure for OpenClaw

Give OpenClaw
a real browser.

OpenClaw agents browse the web to get things done. Local browsers get blocked, fingerprinted, and rate-limited. Spectrae gives your agent managed, stealth-capable cloud browsers — via a single WebSocket URL, with no setup.

Get Early AccessView pricing →

< 1s

Session Start

$0.005

Per Session

2

Browser Engines

∞

Concurrent Agents

01 — How It Works

Three steps.
Running in minutes.

01

Add your Spectrae key

Set SPECTRAE_KEY in your OpenClaw environment. That's the only config you need.

02

Drop in the skill file

Copy browse.py into your OpenClaw skills directory. Your agent can now call browse_page() or screenshot_page() on any URL.

03

Browse any site — undetected

Spectrae's Camoufox engine spoofs canvas, WebGL, fonts, and geolocation. No bot detection, no CAPTCHAs, no IP blocks.

02 — Skill Code

One file.
Full web access.

Drop browse.py into your OpenClaw skills directory. Your agent immediately gains the ability to browse any URL, extract content, and capture screenshots — through an undetectable stealth browser in the cloud.

Standard Playwright API — no new SDK to learn
Camoufox stealth: evades Cloudflare, DataDome, PerimeterX
Switch to Chromium with a single endpoint change
Sessions auto-close — no cleanup code needed
skills/browse.pyCamoufox
# skills/browse.py — Drop into your OpenClaw install

from playwright.sync_api import sync_playwright
import os

SPECTRAE_WSS = (
    "wss://api.spectrae.io/camoufox/playwright"
    + "?key=" + os.environ["SPECTRAE_KEY"]
)

def browse_page(url: str) -> str:
    """
    Browse any URL with a stealth cloud browser.
    Returns the visible text content of the page.
    """
    with sync_playwright() as p:
        browser = p.firefox.connect(SPECTRAE_WSS)
        page = browser.new_page()
        page.goto(url, wait_until="domcontentloaded")
        content = page.inner_text("body")
        browser.close()
        return content

def screenshot_page(url: str) -> bytes:
    """Capture a full-page screenshot."""
    with sync_playwright() as p:
        browser = p.firefox.connect(SPECTRAE_WSS)
        page = browser.new_page()
        page.goto(url, wait_until="networkidle")
        shot = page.screenshot(full_page=True)
        browser.close()
        return shot

Sessions isolated per call

03 — Why Spectrae

Built for agents,
not humans.

No local browser required

OpenClaw doesn't need Playwright, Chromium, or Firefox installed locally. Spectrae handles everything in the cloud.

Stealth by default

Camoufox (Firefox-based) is the default engine — advanced anti-fingerprinting built in. Switch to Chromium for speed-critical tasks.

Concurrent agent sessions

Run hundreds of OpenClaw agents in parallel, each with its own isolated browser session. Infrastructure scales automatically.

PAYG — no commitments

$0.005 per session + $0.002/min. An agent that browses 10 pages a day costs cents per month. No subscriptions.

Ready to connect.

Your agent.
Any website.

Get Early Access

Launching 2026 — PAYG, no minimums