Skip to main content
The models behind the plugin can search the web, read a page you name, parse a document you attach, look at an image, and think through a hard problem before answering. This section covers each one and how to use it. Most of them are already on. There is nothing extra to install and, in most cases, nothing to type.

​
What You Can Do

Reasoning

Pick a reasoning-capable model and it thinks through a problem before answering.

Web Search

The model pulls fresh information from the web when it needs it.

Web Fetch

The model reads a URL you name into its context.

Documents

Attach a PDF or office file with --file and ask about it.

Image Input

Attach a picture with --file and a vision-capable model reads it.

​
What Is On by Default

Web search and web fetch are on by default, on ask, private, session, and fusion alike. The flags exist to turn them off or narrow them, not to switch them on. Reasoning and image input are not flags at all. They are properties of the model you choose: a model either supports them or it does not. Browse the catalog with models to see what each one handles, then pick one that fits the task. The only thing you type every time is --file, because you are naming a specific file to attach.

​
Inside a Session

When you open a full session, your harness’s normal behavior keeps working, now running through Nito. Streaming replies and its own tool use are harness-native rather than plugin flags, so they are covered where the session is.

​
From the Terminal

Every capability here also works from the nito command line tool. See Ask and Fusion.