Alpaca Bot

Easily draft a post or page from any conversation. Dynamically create new content on the fly or with remote resources collected via agents. Alpaca Bot offers a familiar chat interface on both desktop and mobile. You can expect a seamless chat experience on any device!

An Ollama instance is required. Ollama makes it incredibly easy to self-host large language models locally or in the cloud.

If self-hosting isn’t for you, please consider becoming a Patreon for access to our hosted Ollama instances and premium Discord support.

Features

• Chose to store conversation history privately in your wp_ database or not at all.
• Use [alpacabot_agent] to execute tasks on your behalf, generate dynamic content and more.
• Chat with dozens of pre-trained LLMs or train your own.
• Switch conversational model on the fly.
• Create your own custom system messages for highly predictable or formatted responses.

Usage

Text Completion

You have two options to communicate with your AI models;

1. Click Alpaca Bot found in the admin menu, below Dashboard and above Posts.
2. Use the shortcode [alpacabot] to generate a response within any post or page.

Agents

Use the [alpacabot_agent] shortcode to execute tasks on your behalf. Agents are a powerful way to empower your AI models to perform tasks on your behalf.

For example, you can use the [alpacabot_agent] shortcode to retrieve content from a remote source. [alpacabot_agent]s can interact directly with your models and help summarize a webpage or rewrite content.

Example

Basic webpage summarization:

`html 

[alpacabot_agent name=summarize model=tinyllama url=https://example.com/]
`

Shortcodes

[alpacabot] – Chat with Alpaca Bot

Chat with Alpaca Bot from any post or page.

Attributes

• model – The model to use for the text generation. (optional)
• system – Specifies the system message that will be set in the template. (optional)

[alpacabot_agent] – Execute tasks on your behalf

Execute tasks via Agents.

Attributes

The following are core attributes that are supported by all agents.

• name – The agent to execute.

Agent’s communicating with Ollama support [alpacabot] attributes.

Caching

Requests can be cached by setting the cache attribute. cache supports short and long term options.

By default responses are cached to the current post or page.

Transient

Numeric values are treated as seconds and will cache the response for the specified duration.

• cache=60 – Cache the response for 60 seconds.
• cache=3600 – Cache the response for 1 hour.

Post Meta

This is useful for caching responses permanently and associating them with a specific post or page.

• cache=postmeta – Cache to current post or page.

Option

Use WordPress option storage to cache permanently but not associated with a specific post or page.

This can be useful for sharing responses across multiple pages.

• cache=option – Cache to WordPress options.

Disable

The following values can disable caching.

• cache=0 – Disable caching.
• cache=disable – Disable caching.
• cache=false – Disable caching.

Core Agents

The following are core agents that are provided by the Alpaca Bot plugin.

get

Retrieve content from a remote source.

Attributes

• url – The URL to retrieve content from.

summarize

Summarize remote content.

Attributes

• url – The URL to summarize.
• length – Describe the length of the summary.
• content – The type of content we want to summarize.

Support

If you need help or have questions, please join our Discord community.

Premium support and video calls are available to our Patreon subscribers. We can help setup your Ollama instance, troubleshoot issues, demonstrate shortcode functionality and more.

Patreon’s also receive;

• Access to our hosted Ollama instances.
• Priority feature requests.
• Early access to new features and releases.
• Video and community support.

Please consider becoming a Patreon today!

Made Possible By

• Emma Delaney’s How to Create Your Own ChatGPT in HTML CSS and JavaScript
• Google Material Design Icons – Apache-2.0 license
• Hint.css A CSS only tooltip library – MIT license
• TextRank Automatic text summarization for PHP – MIT license
• Ollama Get up and running with large language models locally – MIT license
• Parsedown A better Markdown parser – MIT license