Close Menu
  1. Voices

Voices

Voices alter the way that Nim ‘speaks’, and Nim will use the selected voice while generating content. You can create a voice by entering a prompt, uploading a PDF, or by providing a URL.

For example, if you sell children’s books, craft a voice that’s as friendly and approachable as a puppy at a picnic! Keep the language simple—stick to a 2nd grade reading level or lower—and sprinkle in as many zany words and puns as possible. Make your brand sound like it’s wearing a big, goofy grin!


Screenshot of the main Voices page on NimGPT

When you navigate to the Voices page of your NimGPT dashboard, you’ll see a list of all the voices you’ve created in this project.

If you haven’t created a voice yet, NimGPT will automatically make one for you. Feel free to use this sample voice, it will demonstrate the ways that a Voice can drastically alter generated content.


Creating Voices

You can create a new voice by clicking the Create Voice action button in the top left of the main Voices page.

Creating a new Voice screenshot

Creating a new Voice

Creating a new voice is as easy as clicking the Create Voice button on the Voices page. This will bring you to a new page where you can enter the information required for your voice.

Note: When in doubt, look in the upper right corner of the NimGPT dashboard. Most actions are up there.

Crafting your Voice

NimGPT provides a few different tools to quickly and easily generate a voice that you can immediately use to create content. You can generate a new voice by:

  • Composing a voice prompt directly
  • Referencing a URL
  • Uploading a PDF
  • Generate a voice from text

Composing a Voice prompt directly

If you already know how to write a voice prompt for a LLM, you can dive right in and write your prompt as a voice. Nim will use your prompt when generating content. You can always come back and fine-tune this prompt by editing your voice.


Referencing a URL

You can create a voice from a webpage by referencing it’s URL. What if you already have a well established voice for you content? Simply feed Nim a page that exemplifies that voice.

Nim will crawl the URL and analyze the language used on the webpage, which Nim will formulate into a well-formatted voice prompt. When Nim has finished, Nim will bring you to the Voice editing screen where you can check and see precisely how Nim interpreted the given page into a prompt.

Nim will automatically generate a name for this voice including the URL you used to generate it. We suggest renaming this to be more descriptive of the voice—something you can use to quickly identify it.

Note: Nim is doing a lot of work here, so be patient if it takes awhile while Nim crunches the numbers.


Uploading a PDF

You can create a voice from a webpage by uploading a PDF. Perhaps you have a user manual written a style you like, and want to use this voice to generate more manuals in a similar voice.

Nim will read the PDF and analyze the language used to create it. After reading the PDF, Nim creates a prompt and brings you to the Voice editing screen. You can double check the prompt Nim has generated here.

Note: If you upload a large PDF here, expect this to take some time. Make sure the text of your PDF is machine-readable.


Generate a Voice from text

If you have the perfect example of the voice you’d like to emulate, sometimes the best thing to do is to just use that directly to generate a voice. Nim will read the text you enter and generate a voice from it (much like generating a voice from a PDF or a URL).

A note on how Nim analyzes data to generate prompts

Nim generates voices by analyzing the style of the content you provide, not the style that the example describes. For example, if you upload a PDF of The Elements of Style, Nim will generate a voice that mimics the scholarly writing of that book, not the style of writing that the book encourages. This means that you should provide Nim with examples of the voice in action. Give Nim a blog post written in the voice you want, not the brand book describing the voice.


Voice Fields

Name

This is for your internal use to reference when selecting which voice to use.

Voice Prompt

This is the prompt that gets sent to Nim to guide content generation. It describes the voice and gives Nim parameters to use.

Webpage URL

When generating a voice from a URL, this is the single webpage Nim will crawl to generate the voice.

PDF

When generating a voice from a PDF, Nim will use the content of this PDF to generate a voice.

Text

When generating a voice from a text, Nim will use this text to generate a voice.

Positive Words

These words will be emphasized when using this voice. Nim will go out of its way to use these words.

Positive Words

These words will be avoided when using this voice. Nim will go out of its way to avoid the use of these words.


Using Voices

Whenever you generate content, Nim allows you to specify a voice.

Using a voice when using a shortcut.

Use your Voice!

Whenever you generate content, you can attach a Voice to the generation prompt. Nim will generate different results for the same request just by using a different Voice. For example, say your organization has different voices for marketing and technical documentation, you can just create and use these two voices wherever applicable.

Project settings screenshot

Setting a default Voice

You can set a default voice in your project settings, which will be loaded by default any time that Nim prompts you to use a Voice. If you’d rather select a Voice manually every time, feel free to leave this blank.


Updating Voices

Your voice is a powerful tool, and you should be sure to tailor it to your needs as they change.

The edit voice screen.

Don’t stop tweaking!

You don’t have to be satisfied with the first voice or generated voice that Nim creates. We encourage you to continually improve these prompts as needed! The magic of NimGPT is how granular and personal you can tailor your content generation – so get granular and personal! Tell it what to do and what to avoid if it’s not giving you what you’d like.


Voices