Zhipu AI · Training
ChatGLM-Spider
A `+` URL inside a user agent is supposed to lead to the crawler's own account of itself, which makes the one in circulation for this token worth actually opening: `https://zhipu.ai/spider` serves no crawler policy. It resolves to `www.zhipuai.cn/spider` and renders Zhipu's ordinary marketing homepage (model lineup, news carousel), because the route is a catch-all that returns 200 for arbitrary paths. Zhipu's public presence is model- and product-shaped throughout: GLM releases, AutoGLM, the z.ai and 智谱清言 consumer surfaces, MaaS API documentation, and no webmaster section, bot index or robots guidance anywhere reachable. So the crawl cadence, the scope and the robots.txt behaviour are all unverified, including the much-repeated claim that Zhipu's official documentation commits it to honouring robots.txt, a claim whose cited page is the marketing site.
- Operated by
- Zhipu AI
- Purpose
- Training
- robots.txt token
- ChatGLM-Spider
- Verification
- User agent only
Collects pages into a corpus used to train models.
How to verify ChatGLM-Spider
Zhipu AI publishes no list of IP addresses for ChatGLM-Spider, so nobody (us included) can prove that a request carrying this user agent really came from Zhipu AI. The name is trivial to copy. Treat it as a claim the visitor is making about itself, not as an identity anyone has checked.
There is no range list to confirm. This entry was last reviewed against Zhipu AI's own documentation on .
User agent
Zhipu AI has not published a full user-agent string for ChatGLM-Spider. Requests are identified by the ChatGLM-Spider product token appearing in the User-Agent header; we match that token rather than a whole string, because the rest of the header varies and matching it would miss real traffic.
robots.txt for ChatGLM-Spider
Block
User-agent: ChatGLM-Spider
Disallow: /Allow
User-agent: ChatGLM-Spider
Allow: /robots.txt is a request, not an enforcement mechanism. It is honoured by convention, and a crawler that ignores it is stopped at your edge, not in a text file.
What blocking ChatGLM-Spider costs you
There is no documented surface to be excluded from, so nothing measurable is forfeited. If the token collects training material as third parties assert, the entire trade is whether your text informs a future GLM model: a corpus generates no impressions, issues no citations and sends no referrals, so no click goes missing when the rule lands. What the rule cannot do is bind Zhipu, which has published no commitment to read one. For a site with no Chinese-language readership that makes this close to free; for one that has such readers the cost stays upstream and reputational rather than showing up in analytics.
Vendor documentation
Zhipu AI does not publish documentation for ChatGLM-Spider that we could find. Everything on this page comes from what they do publish elsewhere and from observed behaviour, so treat it accordingly.