Agent Builder
Generates new AI agents from a plain-English description.
Theme
Palette
Your own colour
Saved for this browser, on every page.
Templates
Checked agents, workflows and tools. Read what each one does and what it returns before you install it into your own workspace.
The catalogue
Generates new AI agents from a plain-English description.
Reviews a draft agent spec and returns a polished, production-ready version.
The simplest pipeline: input → agent → output. Edit the prompt template on the agent node to see how variable substitution works.
Routes the run down "true" or "false" based on a tiny expression evaluated against the input.
Loops over an array. The body between iterator and iterator_end runs once per item; results are collected into an array.
Mustache-lite template — every {{key.path}} is resolved against the upstream payload. No code, no LLM call.
Pauses for N milliseconds between two agent calls. Useful for rate-limited APIs.
A trigger node fires this workflow on a cron schedule. The trigger itself is a no-op at runtime.
Fetch any public web page and return it as clean Markdown, with the navigation and scripts stripped. Use this whenever you need the actual contents of a URL rather than a search result summary.
Return the lead summary of an English Wikipedia article by its exact title. Use wikipedia-search first when you are unsure of the title.
Search English Wikipedia and return matching article titles with snippets. Use this to find the exact title to pass to wikipedia-summary.
Turn a place name into latitude and longitude, with country and population. Run this before current-weather, which needs coordinates rather than a name.
Current temperature, wind and conditions for a latitude and longitude, plus a three-day forecast. Get the coordinates from geocode-place first.
Convert an amount between currencies at the European Central Bank reference rate. Returns the converted amount and the rate date.
No template matches that search.
Installing one
Installing starts from registration, so the copy lands in a workspace you own.
It arrives as a draft you can read, rename and change.
Add the connections it needs, try it, then release it.
It arrives as a draft you own. Change it, try it, release it.