waveterm-docs/docs/faq.mdx

---
sidebar_position: 7
id: "faq"
title: "FAQ"
---

# FAQ

### How do I set up my own LLM?

You must manually edit the [config file](./config) located at `~/.waveterm/config/settings.json`.

| Key Name | Type | Function |
|----------|------|----------|
| ai:baseurl | string | Set the AI Base Url (must be OpenAI compatible) |
| ai:apitoken | string | your AI api token |
| ai:name | string | string to display in the Wave AI block header |
| ai:model | string | model name to pass to API |
| ai:maxtokens | int | max tokens to pass to API |
| ai:timeoutms | int | timeout (in milliseconds) for AI calls |

Here's an example of pointing it to a local Ollama instance.  Note that to get the text in the header of the AI block
to update, you'll need to set the "ai:name" key.  For ollama, you'll also need to provide something for the
apitoken (even though it is ignored).

Here are the ollma open AI compatibility docs: https://github.com/ollama/ollama/blob/main/docs/openai.md

```json
{
	"ai:baseurl": "http://localhost:11434/v1",
	"ai:name": "llama3.2",
	"ai:model": "llama3.2",
	"ai:apitoken": "ollama"
}
```

### How can I see the block numbers?

The block numbers will appear when you hold down Ctrl-Shift (and disappear once you release the key combo).

### How do I make a remote connection?

There is a button in the header.  Click the <i className="fa-sharp fa-laptop"/> or <i className="fa-sharp fa-arrow-right-arrow-left"/>
and type the `[user]@[host]` that you wish to connect to.