Tutorial for Contentyze, a content generation platform.

In this post I’ll show you how to use Contentyze to:

  • generate blogposts

  • define and use templates

  • post generated content to Twitter or Wordpress

Contentyze is a powerful content-generation tool, but some of its features might be daunting at first. So let me go step by step to show you that with little training you can do easily extraordinary things, saving a lot of time on creating content.

Step 1: Using Editor or AI-writer

The first thing you see when you enter Contentyze platform is Editor:


The main goal of Editor is to create individual texts using our AI models.

There are four models you can use:

  • General for generating text from a headline,

  • Amazon for generating a review from a link to Amazon product,

  • Rewrite for creating a rewording of an existing text,

  • Summary for summarising an existing text.

General model

This model is used for generating text on any subject, especially when you’re looking to start writing a new text. You shouldn’t expect that the text will be ideal or perfectly logical. The goal here is to unstuck you and generate keywords.

You can control the length of the text by changing "Length" (the number of words to be generated) on the right in Features.

The best way to use it is to input full questions or sentences like:

  • How to become a good writer?

  • How to build a successful business?

  • If you’re interested in real estate, this post will give you an overview of what you should know before starting.

Amazon model

This model is used for generating a review of an electronic product that can be found on Amazon. You simply need to enter a valid link to the Amazon product page. The model works best with:

  • laptops

  • monitors

  • tablets

  • phones

Rewrite model

This model is used to rewrite the existing text. You can either paste the text into a prompt field — it’s extendable so feel free to make it larger — or you can paste a link to a text and our platform will first try to download the content and then rewrite it.

You can choose between abstractive and extractive rewriting by writing in Features ‘true’ or ‘false’ in the abstractive parameter.

  • abstractive — tries to use new words and describe general concepts instead of repeating what’s in articles.

  • extractive — tries to be faithful to the input text; usually doesn’t change any words but make some cuts and joins various sentences.

Summary model

This model is used to summarise the existing text. It works similarly to the Rewrite model: you can either paste your input text or provide a link to an article to be summarised.

You can also choose between abstractive and extractive summarisation, by changing in Features which one you want.

Step 2: Bulk Writer

The second screen you might venture into is Bulk Writer:


The goal of Bulk Writer is to generate content at scale — that means using multiple times the same model for various prompts all at once.

You can use the same models as in Editor, but this time you need to provide a .txt file with links or prompts for a particular model.

For example, you want to summarise 10 texts, you need to provide a .txt file with 10 lines, each line being a link to the article you want to summarise. Alternatively, each line can be the whole text you want to summarise.

The same would work for other models. General model will generate a number of texts, taking each line as a new prompt. This is perfect if you have a list of sentences you want to use — one sentence being one initialisation of General model.

After you import the text file, you’ll be able to generate texts — the process might take even a couple of hours, but feel free to close your browser. You’ll be able to download a .zip file with all the texts from this Bulk Writer page when they are ready.

Step 3: Templates

Now we’re getting into more advanced features — templates:


The goal of templates is to give you full control over generated text. You are able here to mix various models, add your own functions and data, or just script every text.


First of all, you should choose the formatting. You can either choose:

  • General if you want to download generated texts to your local computer.

  • Wordpress if you want to post generated texts to your Wordpress site.

  • Twitter if you want to post generated texts to your Twitter channel.


Secondly, you need to import data on which your texts will be based. That can be:

  • a .csv file or an Excel spreadsheet with various specifications

  • a .txt file

  • an RSS feed you want to use

For example, I can import the following simple .csv file:


As you can see there are 4 rows, the first one being a general "Brand" and “Product” and then there are various examples.

Now the first checkbox asks you whether you want to treat the first row as variables and be able to call them via [[Brand]] and [[Product]] in our case. This is a default option, but if you have just data without descriptions you might want to uncheck.

You can always refer to particular column by [[row_1]], [[row_2]], [[row_3]] and cetera; that is [[Brand]] will give you the same effect as [[row_1]] in our case, and [[Product]] as [[row_2]].

Let me know explain how you use these variables.

I could write in the field below "[[Brand]] produces [[Product]]" and then if I click on Preview below, Brand and Product will be automatically taken from a random row from your data. In my case, I got “Samsung produces phone”:


This way you can create fairly complex templates for generating narratives directly from your data. This can help you create templates for things like:

  • ecommerce descriptions

  • traffic or finance news

If you want to download all the generated texts (a template initiated with each row from your data), then simply click on Download button below Preview. You’ll be able to download the .zip file with every text.

Models in Templates

Let’s go now even deeper into templates. The amazing thing about Contentyze is that you can start adding various models to your templates as well, thus creating more variance in your text.

These are the same models you already know: General, Amazon, Rewrite, Summary. You can add them by clicking on Models dropdown menu and choosing which model you want to apply. Then the text will appear in the form of {%Model "Text…", parameters %}. You can ignore the parameters part and just think about “Text…” — like before in Editor, this is where you enter your prompt: sentence, question, link, longer text.

For example, I entered the following (using data as before):


{% General "[[Brand]] produces [[Product]]", length=250, temp=0.7, top_k=40, top_p=1 %}

What this line does is, it initializes [[Brand]] and [[Product]] first — in our case "Samsung" and “phone”, and then passes it to General model which generates a text-based on prompt “Samsung produces phone”. If we click on Preview we see an example of a generated text:


Saving and Downloading

If I would click right now on Download, I would receive a .zip file with 3 texts, each one being partly taken from the template and partly from a generative model.

Finally to save a template for later use, just click on File->Save or File->Save As and save your template. This is important for Automation later on or just to reuse the same template for a different dataset.

All in all, templates are extremely useful if you need to generate content at scale and take full control over what is generated.

You can also follow our video tutorial on templates here:


Step 4: Automation

The last bit of Contentyze I want to discuss is Automation:


The goal of Automation is to allow you to post your generated texts directly to Twitter or Wordpress — we plan more integrations in the near future.

This is straightforward:

  1. Enter the name of your automation flow

  2. Choose Data you want to use to generate your texts (prompts, links, texts, RSS feed, etc.)

  3. Choose a Template from which you want to generate texts. If you don’t have a template yet, just go to Templates page, create your template as we have discussed in Step 3 and click File->Save or Save As to be able to access it later on here in Automation. Remember to choose Wordpress formatting if you want to post the generated text on Wordpress later on.

  4. Enter your Twitter or Wordpress details to indicate where generated texts should go.

  5. Choose Trigger — the default is "On Data" which means the texts will be posted to your channel as soon as they are generated.

  6. Save automation if you want to use it later on — with different data or templates but the same details for Twitter/Wordpress.

  7. Run automation to make everything go. Now you just have to wait to see results on your channel.

Contentyze tutorial

And that’s it for a tutorial to Contentyze.

You are now able to create pretty complex Automation flows with content. Feel free to experiment as you like and reach out to us if you have any feedback. Contentyze is still in beta and we’re actively developing the project.

You can sign up directly to Contentyze here.

Visit our blog to read more about Contentyze.