liveBook · Manning

Plugin

As explained in LLMs in Production

OpenAI Plugins are extensions that let ChatGPT access third-party tools and external APIs.

OpenAI Plugins let third parties connect tools and external APIs to ChatGPT. They address the appeal of giving an LLM capabilities beyond generated text, but add prompt, token, reliability, and implementation costs. section 7.4.1

A plugin integration typically requires:

  • An OpenAPI configuration file describing and standardizing the API. section 7.4.1
  • An ai-plugin.json file defining the plugin name, authentication, and descriptions used to prompt ChatGPT. section 7.4.1
  • Hosting both files where the API exists, then registering the plugin in ChatGPT for review. section 7.4.1

Practical trade-off

A plugin could expose a weather API to ChatGPT, but an LLM might choose the wrong tool, such as web search, or fail to use a tool at all. Plugin development therefore requires both a clean tool and an effective prompt. section 7.4.1

How LLMs in Production covers it

How deep it goes
Explains configuration, hosting, registration, review, and operational limitations without detailing API implementation.
Examples
Uses weather lookup, web search, OpenAPI configuration, and ai-plugin.json as concrete integration examples.
What you can do after
Build it from scratch by describing an API, defining plugin metadata, hosting files, and submitting the integration for review.
How it connects
Links plugins to LLMs, tool use, prompt engineering, OpenAPI, authentication, and external APIs.
Cover of LLMs in Production
liveBookmanning.com
Read LLMs in Production on liveBook Get the book at manning.com

Also explained in

Compare how 6 books explain plugin →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of LLMs in Production. The section links point to the parts of the book it draws on.