The Talven CX Studio crawler

You are probably here because you found this in your access logs:

TalvenCXStudioBot/1.0 (+https://talventech.com/crawler; commissioned by the site owner)

This page answers, in order: who we are, why we were on your site, how to verify that, how to slow us down or stop us, how to keep us out of your analytics, and who to write to.

Who we are

Talven CX Studio builds customer-facing assistants for businesses, from the content of their own websites. The crawler reads a site to build that site's owner's knowledge base. It is operated by Talven (talventech.com).

Why we were on your site

We only crawl sites whose owner engaged us. A business signs up, names their own domain, and the crawler reads it. Nobody is crawled for having been noticed, and we do not crawl the open web. If you run the site, someone in your organization commissioned this. If you believe nobody did, write to us — that is exactly the mail we want to receive.

How the crawler identifies itself

Plain HTTP fetches (robots.txt, sitemaps, availability probes) send the user agent above plus a From: [email protected] header — the standards-defined address of the human responsible for the requests (RFC 9110 §10.1.2).

Page rendering uses a real Chromium browser, and its user agent is the browser's own with the same token appended:

Mozilla/5.0 (...) Chrome/<version> Safari/537.36 TalvenCXStudioBot/1.0 (+https://talventech.com/crawler)

Grep your logs for TalvenCXStudioBot and you see everything we did.

How we behave

• One crawl per site at a time, fleet-wide — never parallel crawls of your server.
• A handful of requests in flight at most, with deliberate delays between them.
• Scheduled re-crawls fetch deltas, not everything, on a cadence the site owner sets.

Keeping the crawler out of your analytics

By default you should see nothing. The crawler blocks analytics collection requests inside its own browser — Google Analytics, Matomo, Plausible, Meta pixel, Segment, and session recorders — so a crawl records no pageviews, no visitors, no sessions. It also sets the Plausible and Matomo opt-out flags. Your content, tag manager, and scripts load normally; only the measurement beacons are stopped.

If you want defense in depth, or you have a vendor we do not block:

Filter by IP (works everywhere). All crawler traffic comes from:

34.65.142.180

(IPv4 only today. The address is a reserved Google Cloud static IP in Zürich; it survives infrastructure changes on our side.)

We treat these addresses as a commitment: changes are announced on this page in advance and old and new run in parallel — never a silent swap.

• Google Analytics 4: Admin → Data collection and modification → Data filters → Create filter → Internal Traffic; match by the IP above (CIDR accepted). Mind two GA4 rules: a property allows at most 10 data filters, and an active exclude filter is permanent and forward-only — it cannot clean up the past.
• Matomo: Administration → Websites → Settings → Global list of excluded IPs; or exclude by user agent with the string TalvenCXStudioBot.
• Plausible: nothing to do — the crawler sets plausible_ignore on itself.

Filter by user agent (server logs, CDNs). Match TalvenCXStudioBot. The token contains Bot, so most bot-classification rules catch it without configuration.

Slowing us down or stopping us

Write to the address below with your domain. Because every crawl belongs to a specific customer of ours, we can slow it, reschedule it, or stop it entirely — and if you are the site owner, your own account controls the crawl schedule directly.

Who to write to

[email protected] — this mailbox is read by the people who operate the crawler. Include your domain and, if you have it, a log line; we can match any request to the crawl that made it.