Tools
web
Web
Read the open web.
Actions
read
read url=<address>
One page as clean text: a company's site, an article, a pricing page, a link someone shared. Pictures are left out.
RESTGET/web/read·CLItracee web read
images
images url=<address> [query]
The pictures on a page (team photos, avatars, product shots, logos), each with its link, alt text and the text right before and after it. A picture's caption is usually the text right after it ("Katie | AI Outbound Agent" after Katie's avatar). query ranks the ones that match first. To use one, add it with media action=add (or brand action=add_asset): the library keeps a copy in a format every network takes.
RESTGET/web/images·CLItracee web images
search
search query=<words> [limit]
Web results with titles, links and snippets; then read the ones worth it.
RESTGET/web/search·CLItracee web search
map
map url=<site> [limit]
Every page address on a site (up to 300), to find its pricing, customers or blog pages before reading them.
RESTGET/web/map·CLItracee web map
Use it to research a company or a topic before writing about it. For the workspace's own brand, brand action=learn saves what the site says.
Inputs
Every call takes action and the fields that action names above.
actionstringrequiredWhat to do: read, images, search or map. Each is described in the tool description.
Available options:
readimagessearchmapurlstringFor read, images and map
up to 500 characters
querystringFor search; for images, what to look for, e.g. "Katie avatar"
up to 300 characters
limitintegersearch: results, default 5 (at most 10). map: pages, default 300
1–300