> ## Documentation Index
> Fetch the complete documentation index at: https://docs.qxlabs.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI nodes

These nodes put a model to work inside your flow, whether that is answering a question, researching a topic, or sorting through a list. They all consume credits, see [Plans and credits](/getting-started/plans-and-credits).

* **Query AI** sends a prompt to a model and hands you the answer. It has no web access, so give it everything it needs in the prompt.
* **Web Research** researches a topic online and returns an answer along with the sources it used. Use this instead of Query AI when the answer depends on current information.
* **Run an Agent** calls one of your [agents](/agents/overview) as a step in the flow, so it arrives with its own instructions, tools, and knowledge.
* **AI Categorizer** sorts a list into buckets you define. You write the categories, and it returns one per item in the same order, so the results line up with what you sent in.
* **AI Filter** keeps only the list items that match a condition you describe in plain language.
* **AI Data Extractor** pulls specific values out of messy text, such as a name and a price from an email. Define a field for each value you want, and each one becomes a separate output for later nodes to use.

<Note>
  Filtering nodes can keep other lists in step with the one being filtered, so rows stay matched up. Add them under **Associated columns** and they are trimmed alongside the main list.
</Note>