Skip to content
Oculr
Web scraping

Collect public web data with real browsers

Oculr gives data teams real Chromium profiles that behave like real browsers, each with its own proxy and a consistent identity, driven by your existing Selenium, Puppeteer and Playwright scripts or an MCP agent.

GoogleRedditYouTube
20+
Pinned Chromium kernels
5
Proxy protocols per profile
Hundreds
Profiles per fleet
Real browsers, at fleet scale

Collect public data the way a real browser would

Real Chromium profiles, one consistent identity each, driven by the tools you already use and run in parallel as a fleet. Public data only, with each site's terms and rate limits respected.

Run a fleet of collectors

Fleet commands create, launch, navigate and act across many profiles at once, each keeping its own proxy, cookies and fingerprint, so you collect across regions and sites in parallel.

Agentvia MCPInstagramDEAmazonUSRedditUKTikTokJPEtsyCA

Every collector a consistent identity

Each profile pins an exact Chromium kernel and a deterministic fingerprint, so a collector presents the same identity on every launch and a run stays reproducible across the whole fleet.

  • Pinned Chromium kernel per profile
  • User agent and client hints match the real engine
  • The same profile, the same identity every launch
  • Reproducible runs across the fleet
Public data only. Respect each site's terms, robots and rate limits.
Why it matters

Collect public data from real, consistent browsers

Generic automation stacks announce themselves and drift between runs. Oculr profiles are real Chromium with aligned, deterministic identities, so collection is consistent and reproducible. Public data only, with each site's terms and rate limits respected.

Drive it from the tools you already use

Launching a profile returns a WebDriver-compatible driver path plus the CDP websocket endpoint, so your existing Selenium, Puppeteer and Playwright scripts attach unchanged. Prefer something lower level or higher level? Speak raw CDP, call the local REST API, or hand collection to an MCP-compatible agent.

  • Selenium, Puppeteer and Playwright attach unchanged
  • Raw CDP and a local REST API underneath
  • Or drive profiles from an MCP-compatible agent
# Launch a profile, get standard endpoints
driver_path, cdp_ws = oculr.launch("price-de-fra")
# Your existing script attaches unchanged
driver = webdriver.Chrome(driver_path)
await playwright.chromium.connect_over_cdp(cdp_ws)
Selenium, Puppeteer, Playwright, raw CDP, REST or MCP

A proxy and a consistent identity per profile

Assign each profile a residential or datacenter proxy over HTTP, HTTPS, SOCKS4, SOCKS5 or SSH, and Oculr aligns timezone, language and geolocation to the exit IP automatically. Connect a provider once from the vault and bind profiles to it instead of pasting URLs.

  • Residential or datacenter proxies per profile
  • Timezone, language and geo align to the proxy exit
  • Provider vault: Bright Data, Oxylabs, Decodo and more
collector identities4 regions
serp-us-east
New York, USAmerica/New_York · en-US
serp-de-fra
Frankfurt, DEEurope/Berlin · de-DE
serp-jp-tok
Tokyo, JPAsia/Tokyo · ja-JP
serp-br-sao
Sao Paulo, BRAmerica/Sao_Paulo · pt-BR
Timezone, language and geo aligned to each proxy exit

Reproducible runs at fleet scale

Pin each profile to an exact Chromium kernel version so the user agent and client hints always match the real engine, and a deterministic fingerprint means the same profile presents the same identity every launch. Fleet commands then run collection across many profiles at once, each with its own proxy and identity.

  • Pin browser versions per profile for reproducible runs
  • The same profile presents the same identity every launch
  • Fleet control for parallel collection across regions
fleet status3 running
Shopifyprice-de-frachrome 146page 4 / 12
eBayprice-uk-lonchrome 147page 9 / 12
AliExpressstock-us-nycchrome 146page 2 / 8
Etsyreviews-jp-tokchrome 148pending
FAQ

Frequently asked questions

What kind of web scraping is Oculr built for?+

Collecting publicly available web data: price monitoring, market research, SERP tracking and competitive monitoring. Oculr gives your collectors real Chromium browsers with consistent, isolated identities. Always respect each site's terms and robots directives and keep your request rates sane.

Will my existing Selenium, Puppeteer or Playwright scripts work?+

Yes. Launching a profile returns a WebDriver-compatible driver path plus the CDP websocket endpoint, so your existing Selenium, Puppeteer and Playwright scripts attach unchanged. You can also speak raw CDP directly, call the local REST API, or drive profiles from an MCP-compatible agent.

How do proxies and geo settings work for collection?+

Assign each profile a residential or datacenter proxy over HTTP, HTTPS, SOCKS4, SOCKS5 or SSH, and Oculr aligns timezone, language and geolocation to the exit IP automatically. Connect a provider once from the built-in vault, including Bright Data, Oxylabs, Decodo, IPRoyal, SOAX, NetNut or a custom provider.

Can I reproduce a run exactly later?+

Pin each profile to an exact Chromium kernel version, from Chrome 86 to the current majors, and the user agent and client hints always match the real engine underneath. Combined with a deterministic per-profile fingerprint, the same profile presents the same identity every launch, so a collection run is reproducible.

Can one process run many collectors in parallel?+

Yes. Fleet commands let you create, launch, navigate and act across many profiles at once, each keeping its own proxy, cookies and fingerprint, so you can collect across regions and sites in parallel.

Built on these features

Collect public data with real browsers

Spin up your first two profiles free, attach a proxy and your existing script, and run a clean, reproducible collection job today.

Start free, 2 profilesRead the docs

No credit card required · Cancel anytime