Your Search engine optimisation information to the ChatGPT API

ChatGPT introduced the rollout of its API (GPT 3.5 Turbo) on March 1. 

I’m bullish on ChatGPT’s utility for a number of completely different Search engine optimisation-related capabilities like key phrase analysis, native Search engine optimisation, content material, and hyperlink constructing. Having spent a lot time utilizing ChatGPT, I’m additionally painfully conscious of its limitations.

Whereas the API received’t be a panacea (and the online interface is definitely a lot better for some duties), it could possibly assist tackle among the shortcomings of the online interface.

This text will present you learn how to maximize ChatGPT’s API with particular Search engine optimisation use instances.

How one can use the ChatGPT API

To leverage the ChatGPT API, you want to have the ability to entry the API within the first place. ChatGPT’s father or mother firm OpenAI has intensive documentation for utilizing the API.

In the event you’re trying to study extra about constructing a software or interacting instantly with the API there’s additionally a great walk-through right here.

You can too use AppsScript to question the ChatGPT API in Google Sheets, which we’ll stroll by right here step-by-step.

No matter your method, you’ll want to start out by getting an API key.

Getting your ChatGPT API key

Upon getting an OpenAI account, you may generate your API key both by following this hyperlink whereas logged in or clicking View API keys within the profile dropdown:

Then click on Create new secret key.

Copy the API key you generated.

Connecting the ChatGPT API to Google Sheets

There’s a simple solution to join ChatGPT to Google Sheets (h/t Keith Mint).

First, create a brand new Google Sheet, then click on on Extensions and select Apps Script:

You then paste the next code (along with your API key rather than YOUR API KEY):

const SECRET_KEY = "YOUR API KEY";
const MAX_TOKENS = 800;
const TEMPERATURE = 0.9;

perform AI_ChatGPT(immediate, temperature = 0.4, mannequin = "gpt-3.5-turbo") {
  const url = "https://api.openai.com/v1/chat/completions";
  const payload = {
    mannequin: mannequin,
    messages: [
      { role: "system", content: "You are a helpful assistant." },
      { role: "user", content: prompt },
    ],
    temperature: TEMPERATURE,
    max_tokens: MAX_TOKENS,
  };
  const choices = {
    contentType: "utility/json",
    headers: { Authorization: "Bearer " + SECRET_KEY },
    payload: JSON.stringify(payload),
  };
  const res = JSON.parse(UrlFetchApp.fetch(url, choices).getContentText());
  return res.decisions[0].message.content material.trim();
}

Click on save inside Apps Script:

Then you should utilize the next perform format to use a immediate to a cell:

=AI_ChatGPT("YOUR PROMPT HERE")

As Mint factors out in his article, you may as well reference a cell if you wish to have a number of cells use prompts that refer again to at least one cell (just like the title or subject of a weblog submit):

Within the instance above, I used easy prompts just like the one pictured, then created a second paragraph for this subject. (We’ll stroll by extra particular functions for the API shortly.) 

ChatGPT API pricing

Earlier than you begin leveraging the ChatGPT API for Search engine optimisation duties, it’s important to grasp the pricing.

The worth for the gpt-3.5-turbo API (the ChatGPT API) is $0.002 per 1,000 tokens, which is 10x cheaper than the present GPT-3.5 API.

To higher perceive what the pricing truly winds up wanting like, you could perceive how tokens work.

ChatGPT API tokens

OpenAI has a good breakdown and a useful free tokenizer software that can assist you decide what number of tokens a selected textual content will likely be and the way the textual content is damaged down (in case you could cut back the variety of tokens for a immediate or response).

Just a few key issues to bear in mind:

• By default, the API can return a most of 4,096 tokens.
• Tokens are a illustration of how a lot textual content your immediate and response are. It is a key issue, as longer prompts can shorten your response output.
• Textual content is translated into tokens and roughly breaks all the way down to round 4 characters in English.

OpenAI additionally offered this beneficial breakdown of how tokens are calculated from textual content:

• 1 token ~= 4 chars in English
• 1 token ~= ¾ phrases
• 100 tokens ~= 75 phrases

Or

• 1-2 sentence ~= 30 tokens
• 1 paragraph ~= 100 tokens
• 1,500 phrases ~= 2048 tokens

To get further context on how tokens stack up, think about this:

• Wayne Gretzky’s quote, “You miss 100% of the pictures you don’t take,” accommodates 11 tokens.
• OpenAI’s constitution accommodates 476 tokens.
• The transcript of the U.S. Declaration of Independence accommodates 1,695 tokens.

So should you used a brief immediate to generate a 1,500-word article, it could be lower than half a cent.

Particular API use instances for Search engine optimisation

The API might help with a number of the identical Search engine optimisation-related duties because the ChatGPT internet interface, however the two have some vital variations, making some duties higher for one than the opposite.

Understanding that may enable you decide which to make use of for Search engine optimisation duties.

ChatGPT API vs. internet interface

The most important variations embrace the next:

Scale and bulk duties

If you wish to combine ChatGPT with an present utility or spreadsheet, you will want to make use of the API. 

Moreover, the API will likely be your good friend if you wish to carry out a selected perform utilized throughout a number of situations (e.g., generate meta descriptions for a number of key phrases).

High quality-tuning and relationship between prompts

Presently, ChatGPT’s API doesn’t assist fine-tuning. In the event you create a number of prompts by the API, they will not have a relationship. 

You may create a system message to use to a number of prompts and responses, however OpenAI has mentioned that these are regularly presently ignored by the gpt-3.5-turbo API. 

This implies when you’ve got a process that requires a number of prompts and for ChatGPT to maintain the context of a complete dialog, you will wish to use the online interface.

Character limits

The API has a token restrict of 4,096 which interprets to round 16,384 characters per interplay for every immediate and response. 

Character limits for the online interface can range, however usually, the immediate and response are restricted to round 2,048 characters, or 4,096. 

So for duties that require extra intensive prompts or extra vital responses, the API will likely be a greater choice. 

There are additionally extra choices for structuring prompts and duties in a approach that offers you a lengthier mixed output should you’re utilizing code reasonably than one thing just like the Google Sheets integration.

Pricing 

Once more, the API is priced by utilization (and presents a free trial with $18 value of tokens), and the online interface presently presents a free model and a $20/mo. paid model. 

Instance ChatGPT API prompts

Let’s take a look at particular use instances the place you’d wish to think about the API over the online interface.

Title tags and meta descriptions

An apparent use case the place the API is sensible is having ChatGPT generate title tags or meta descriptions at scale.

You should use comparable prompts to people who work within the internet interface right here, however should you construction them correctly and lay out your spreadsheet proper, you may take the primary immediate after which rapidly apply that to a number of URLs or key phrases: 

Notice that the free plan significantly is charge restricted, so you will have errors the place the cells outing, during which case you could delete and re-paste:

And as at all times, needless to say:

• ChatGPT cannot crawl the online, so it is both utilizing prior data of a URL from the coaching interval or an assumption concerning the URL based mostly on phrases within the URL to generate solutions
• ChatGPT’s output can typically be flawed or deceptive and must be edited.

You should use this identical format for title tags. (I will have an article focusing particularly on utilizing ChatGPT to generate and replace title tags popping out quickly).

Longer content material and submit outlines

You should use the online interface in ChatGPT to create chunks of content material and descriptions fairly simply.

If you wish to make an extended article or if you are going to create a sequence of outlines on completely different subjects, nevertheless, it may be a clunky expertise.

The API is smoother for these duties.

First, you may chunk off a submit you wrote into sections. Then line up the sections and pull your immediate down:

Alternatively, you might have ChatGPT generate a number of outlines that you just then have a author (or writers) fill in:

Then, you might have the API write the content material one part at a time:

Once more, right here you might take these prompts, load them one after the other throughout the define (altering every immediate for the suitable part), after which simply pull the identical formulation throughout a number of outlines and have a number of textual content particular to the subsections of every submit generated.

My expertise is that this helps you avoid token limits, and even pedestrian prompts just like the one above mixed with having ChatGPT create an overview will generate higher content material than should you ask the software to “write a weblog submit.” 

Past that, you may additional enhance content material high quality by doing issues like:

• Ask it to incorporate particular phrases (both since you need them on the web page otherwise you wish to add inside hyperlinks for that phrase).
• Feed it statistics or quotes to work into particular sections of the submit. (That is significantly useful when you’ve got a subject that requires up-to-date ways and statistics, on condition that GPT-3 was not skilled on current knowledge.)
• Tweak your prompts to output the tone and formatting you need (extra on this shortly).

And, as at all times, layer on human modifying.

FAQs

There are a number of FAQ-related capabilities the API might help with.

First off, you may generate an inventory of FAQ inquiries to be related to a weblog submit:

Subsequent, you may have the ChatGPT API reply these (once more: proceed with warning in the case of output high quality and accuracy):

Schema

You can too have ChatGPT generate schemas for you throughout a number of pages. 

On this case, we will apply FAQ schema to the FAQs it created for us:

Convert content material to HTML

One other cool use case for ChatGPT is to transform textual content to HTML. 

First, let’s convert our submit from textual content to HTML:

Just a few issues to notice right here:

• As you may see, the headers in every part had been initially formatted with paragraph tags. I fastened this within the immediate by including, “Format the header of the part as an h2, and every other headlines on this part as an h3.”
• I wasn’t capable of mix your entire submit (which was round 1100 phrases) into one cell to be transformed to HTML, so I needed to chunk it off and gave particular directions in my immediate for every cell to verify ChatGPT did not layer within the formatting for a complete HTML doc in each cell.
• You might also get some wonky formatting, like further quotes you did not anticipate within the output. 

As with all elements of ChatGPT, hold refining your prompts and at all times test the output.

Add inside hyperlinks

Now that we’ve got our submit in HTML, we will begin to do some cool issues with inside linking.

We are able to inform ChatGPT so as to add inside hyperlinks from particular phrases to particular pages wherever within the HTML we simply generated:

If we had been making a cluster of pages, we might create guidelines particular to every web page and apply them to the related HTML.

This fashion, all the pieces we generate is interlinked the way in which we wish, the HTML is able to go, and the FAQ schema is added appropriately.

I attempted to create a rule that was somewhat extra difficult, asking ChatGPT to hyperlink phrases. Here’s what I added to my immediate:

• “Any time you see the phrases making and snicker inside 5 phrases of one another, create a hyperlink with these two phrases and all the phrases in between them to standupexperience.com/make-them-laugh.”

Right here is the hyperlink that was added:

That is not what I requested for, and in addition, it did not hyperlink each occasion of stand-up as I requested it to.

I will have to refine my prompts and test my outputs.

Outreach templates

In the event you’re reaching out to a number of locations for hyperlink outreach or visitor submit placements, you should utilize the ChatGPT API to create a number of outreach templates for you:

You probably have completely different sorts of templates, get inventive with making use of these prompts throughout completely different outreach targets.

Utilizing the online interface and the API in tandem (a.ok.a., deliver your prompts)

Two issues are true:

• The API is healthier for bigger initiatives and performs many prompts concurrently.
• The online interface is regularly higher at getting you to a wonderful output since you may trip with ChatGPT to provide context, assessment the output, present suggestions, and so forth.

One solution to get the most effective of each worlds is to create your immediate within the internet interface after which apply it to a number of gadgets by way of the API.

Let’s take a look at a selected instance from this submit. My inside linking immediate did not work by way of the API. It is difficult to troubleshoot by way of the API as a result of you may’t give ChatGPT suggestions or ask why the immediate failed.

Enter, immediate era by way of the online!

My typical ChatGPT course of is to: 

• Give it chunks of context.
• Verify for understanding. 
• Overview the output for errors and provides it further suggestions to attain my desired consequence.

To have the ability to do that for inside linking guidelines, let’s begin with the online interface to scrub up our ChatGPT immediate.

First, I wish to give ChatGPT some context. (Keep in mind: ChatGPT does not find out about its personal API!) 

I wish to give it details about the API, share the HTML I am beginning with, share the immediate I used, and share the output I acquired after which ask ChatGPT to diagnose the problem and rewrite the immediate for me.

Let’s examine the way it goes:

In the event you do not pre-empt it, ChatGPT will repeatedly interrupt you to fireplace (regularly irrelevant) solutions and solutions (like an keen scholar who did not do the studying).

I copied and pasted the textual content from this web page in 4 prompts, the HTML I used to be making an attempt so as to add hyperlinks to, the total Sheets perform I would used, and the output I would gotten.

Subsequent, I lastly shared the problem with the output:

After which ChatGPT fastened the immediate for me.

Properly, it wasn’t truly fastened.

ChatGPT is relentlessly well mannered even should you’re getting a bit testy, and once I shared the immediate, it analyzed it for me:

The solutions on HTML measurement are good ones, however I used to be nonetheless getting the error:

This time, the immediate labored!

To handle the max tokens problem, I can go to Apps Script to regulate that quantity:

Or clearly, if I am not utilizing Sheets, it is not more likely to be a problem (till I hit the 4096 tokens).

Get inventive and search for options

Search engine optimisation functions for the ChatGPT API go properly past what’s listed right here. 

In the event you’re looking out for methods to make use of the platform and might get inventive you will discover many extra functions like:

• Programmatic options: You may transcend Google Sheets to seek out thrilling and scalable makes use of for the ChatGPT API.
• Combining a number of APIs: Take into consideration the way you may have the ability to use the ChatGPT API in unison with APIs like Google Search Console, Ahrefs, Semrush or comparable. What issues do you wish to resolve? 
• Extra environment friendly or higher workflows: Take inventory of the duties you and your staff full every day. Are there gadgets that ChatGPT’s API might be able to both will let you just do as properly however extra effectively, or to enhance your work product round?

With the current launch of GPT-4, extra alternatives for leveraging the platform will possible proceed to crop up.