Laurseraph

Technique
Credential phishing
Sub-technique
BitM
Targeted platforms
Google
Also tracked as
GhostFrame

Updated · First published

Current breakthrough rate

Moderate

  1. Apr 2026: Moderate
  2. May 2026: Moderate
  3. Jun 2026: Moderate
  4. Jul 2026: Moderate
  5. Aug 2026: Moderate
  6. Sep 2026: Moderate

Laurseraph is a browser in the middle phishing tool targeting Google accounts that emerged in the second half of 2025. It had a small surge of activity at the start of 2026 before becoming a more established phishing kit throughout summer 2026.

Laurseraph phishing kit fronted by two fake CAPTCHAs

The browser in the middle mechanism works by creating a new cloud browser session on the attacker's infrastructure that connects to the Google login form. Images from this browser are displayed to the victim in a canvas on the phishing site. Interacting with this canvas relays the victim's mouse clicks and keystrokes back to the attacker's browser.

URL patterns

The main URL that is visible to recipient is that of the anti-bot gate. This has most frequently been hosted on a Cloudflare Workers domain. Push has observed two main naming patterns during 2026:

  • accounts-g0033le-com-[4].[tenant].workers.dev
    • Example: accounts-g0033le-com-1cd3.laurseraph.workers.dev, accounts-g0033le-com-f753.vn604074.workers.dev
  • [word]-[word]-[4].w-[6].workers.dev
    • Examples: session-base-paperdoc-k093.w-jdfq64.workers.dev, secure-proposal-l4f5.w-kkjhaf.workers.dev

The browser in the middle phishing client page is loaded as a Blob URL with connections to the remote browser relay server being performed in the background over a WebSocket:

Plaintext
blob:https://silverpoints-ehax.w-jo6fbr.workers.dev/1de51af1-3cd1-4e7b-8260-29a0976af281

Network traffic

After passing the anti-bot gate, the kit sends a liveness check request to the /eal endpoint. In some deployments this was sent to a list of different central C2 servers with the first to respond successfully picked as the C2 for the follow up session creation request.

A new phishing browser session is created with a POST request to the /create endpoint:

HTTP
POST /create HTTP/1.1
Content-Type: text/plain

{"hash":"307YyWxq","ttl":60,"e":"","h":"","c":"","r":"","t":"0","bl":"2","op":"0","domain":"invitation-to-bid-4xbi.w-f859x0.workers.dev"}

{"success":true,"html":"[300kb base64]","html_encoded":true,"mode":3,"mode_url":"https://accounts.google.com"}

The HTML in the response contains all of the configuration for connecting to the relay host and receiving images from the remote browser. The connection is established over a WebSocket connection to wss://<relay-host>/ws.

Representative samples

Sample URLs

  1. hxxps://proposal-cl5y[.]w-zeozn7[.]workers[.]dev/

    urlscan
  2. hxxps://base-page-bb2x[.]w-nloxx7[.]workers[.]dev/

    urlscan
  3. hxxps://view-shared-document-taaf[.]sundrese[.]workers[.]dev/

    urlscan
  4. hxxps://accounts-g0033le-com-1cd3[.]laurseraph[.]workers[.]dev/

    urlscan

Public analysis

Third-party research on this tool.

See more

TechniqueSub-techniqueTools
BitM
  • Laurseraph
Authorization phishingDevice code phishing
Malicious copy & pasteClickFix