Intelligent Rewrite

This guide offers an overview of using Cephable Rewrite, enabling users to efficiently rewrite content while maintaining privacy and utilizing on-device AI. It provides instructions for making text more concise, including examples for emails and code, and highlights the convenience of voice controls for hands-free operation. By following this guide, users can enhance their writing process with ease and flexibility, without compromising security.

Cordellia
|
17 steps
|
37 seconds

Cephable Rewrite allows users to use LLM rewrite features and iterate on content all while staying Private, Offline, and using the capabilities of your device (On-Device AI).

Screenshot of the Cephable Desktop App with the Panel Window. The Click Target is on Intelligent Tasks.

Let's look at some examples. This is an email to congratulate the team on the latest release. It is wordier than we would like however, and we would like to make it more concise.

An outlook email window with some body content congratulating everyone on a successful release. It highlights how the latest release better fits client needs, was a team effort, and asks members to post on LinkedIn regarding the accomplishment.

Highlight the text you want to rewrite. Then, click the Rewrite Text action in the Intelligent Tasks Section of the Panel.

The email body content is highlighted. And the click target is on Rewrite Text In the Intelligent Tasks Section.

We will use Highlighted Text option for In-App Text Replacement. However, you can also use Paste Text option to copy the highlighted text, and generate content in a new app or location without replacing the previous content.

Click Target in the Describe Text Box for Instructions. And arrow pointing to Paste Text option with a caption "To Rewrite Text in a different location from highlighted text"

Add rewrite text instructions. For example if you want to make the email content more concise, the instructions could be "Make this Concise."

Click Target on instructions with some text "Make this concise."

Click Generate to Start Rewriting. Remember that then Generative AI is all running offline, private, and on-device.

When the content generates, be sure to stay in the app you want the generation / text replacement to occur. Note, when the model is preparing and generating text, the border of the panel changes colors an animated colorful border and the Generate button gets a loading indicator.

The border color of the panel has changed to an animated colorful gradient. And the generate button has a loading indicator when the model is generating content.

Notice that the replacement happens In-App and now the email content is concise!

A rectangle around the rewritten concise email content. Now it is 2 sentences instead of 2 paragraphs.

Let's look at another example with code! The sample has a TypeScript service for managing insurances for clients. The insurance wants to add insurance for Long Term Disability and Life. Let's use Cephable Rewrite for this.

A code snippet of insurance service. The insurance provides health, auto, and home insurance.

We can also rewrite highlighted text in-app using Always-On Voice Controls. Turn on the microphone.

The click target is on the microphone toggle in the panel.

Highlight the text you want to rewrite.

The insurance types are currently highlighted.

Now you can use Voice Controls for your rewrite instructions. Example: "Rewrite to include life insurance and disability".

The microphone voice content has the same content that is being said as an example.

After the voice control is given, a generating loading indicator renders and the code will be rewritten.

The insurance types are still highlighted. And now there is a generating loading state indicator in place of the voice output while the model is working to add the 2 additional insurances.

After the model runs, there will be rewritten content in place of the original text content.

A rectangle around the new insurance types highlights that Life and Long Term Disability Insurance has been added.

You can always adjust your generation settings through the Text Generation Settings. First, in the main dashboard window, click Settings for your Running Device.

The screenshot is back on the main dashboard and the panel on the side.

Click on the Text Generation Settings Button.

The Text Generation Settings in the Device Settings Page is focused.

If you have a tone, style, and other instructions you want to be respected on each Generate, Rewrite, Translate, and more Intelligent Task related actions - this is the place to add additional system prompts. Other things like auto warm up, clearing resources, and cache can also be managed here.

The text generation settings has many configurable settings to improve the user's experience with offline, private AI text generation features.