People are desperate to have their plugins appear on the front page

People are desperate to have their plugins appear on the front page.
They are putting “A” in front of the plugin name so that it appears at the front when sorted alphabetically XD

I’m not trying to criticize this, but isn’t it somewhat amusing?


Next: SEO for the plug-in store search function.
Also, publishing slightly different versions of the same plug-in under different names and accounts etc… pp…
Indeed amusing to see yet another return of this old competition in the plug-in store.


old competition

Has this kind of competition occurred elsewhere as well?

In your opinion, how should it be done?

What would be a fair system for that coveted first page spot?

The situation you posted about is very much comparable to all the tactics we used in the earlier days of the internet where someone would find a way to rank high in a directory/search engine and then all the other marketers piling on trying to get a easy piece of the cake until practices were somewhat devalued when generating search results or ordered lists for the user.
Then some other tactic pops up and the game repeats.
Looking back a lot of this is now somewhat told in a derogatory tone, like wild west black hat SEO, but at the end of the day we all try to push the fruits of our work to the market. And if this means my app is called “AAA AI Chatbot” so be it.


I believe they should be sorted by the number of downloads. However, to give opportunities to new plugins, some pages might need to randomly showcase new plugins.

Very much agree, @vb. I’m not sure I like the results, but if you give clever people an opportunity they will optimise for a best (for them) result case very quickly.

I think the important thing here is giving both the end user a pleasant and enjoyable experience, one which they want to embrace as it genuinely helps them save time/money or gives them entertainment, and it works for developers who want to showcase their offerings, all the while minimising the number of spam, low effort entries.

It certainly seems to be something that needs some form of curation and management.

I wonder if a model (GPT-4 100k?) could be used to summarize all the plug-ins and recommend one or a few based on a query explaining what you want to do.

Text to SERP?


It may be a hard problem for an average user, but it should be very easy for Open AI. The key problem right now is that the plugin database is not officially open to the general public. Open AI could do two things easily:

  1. Vectorize the plugin database. That basically makes the semantic search possible.
  2. Expand their “Custom Instructions” to allow the user to specify what should be on the first few pages.
  3. Voila…the first page will best match the wishes on your mind and the desires in your heart.

That’s more or less what we have done with OhMyPlug plugin for plugins recommendations.
You can try for example
- plugins to retrieve fresh news from the web
- plugins to help me organizing my trip in Korea
- The 2 most relevants plugins to draw diagrams within chatGPT
- Give me more info on plugin “Eraser”
- …

You can use OhMyPlug installing it from the store or directly using this shared conversation :

1 Like

The last two suggestions are great! Definitely an improvement over a simple alphabetical list
Let’s assume I know that a typical search query for a pdf integrator plug-in would be “read from pdf” or the like, then someone else will describe their plug-in as “Best solution for PDFs. Chat and read PDFs. PDF app helps to…”. Additionally, you could reference “pdf” in your manifest file and function names and even spin up a custom website just for this purpose that will be linked to from the store.
People will adapt their plug-in names and descriptions to appear at the top for most people.

This is not a simple challenge to tackle but right now, almost everything would be a improvement over the status quo. If I recall correctly the big search engines use a semi-automated process for the top queries, vetting the first page results manually every so often.

Made a graph of the problem :smile:

I’m following weekly launches of plugins and can see the quality has been going down lately… less big companies making plugins and a lot of the same use cases.

I think the success of ChatGPT plugins really depends on OpenAI implementing better quality assurance at this point.

There’s some really good use cases for plugins, but unless devs have a clear incentive to make them (should be easily monetizable, like iOS apps), and users have an easy way to find actual useful ones for their needs, it’ll remain a stagnating feature.


Additional points awarded if you used Code Interpreter to generate the image!


Great intuition, I absolutely did!


Phone book days it was done e.g, “AAA automotive”:}


Hi and welcome to the developer forum!

All of these non technical issues will be addressed in time, there is a limited team of people at OpenAI right now with an ever expanding userbase. Priorities seem firmly focused on model performance and scaling up access over the finer points of the UX, I’ll admit that there is lots of room for refinements, but it does work and something is better than nothing when it comes to bleeding edge tech. I know they are actively hiring for capable people al the time.


This reminds me of the old days when I’d search for games on Everything was such random quality, often you’d get junk but there were some gems.

It is quite exciting to kind of be in this realm again where things aren’t SEOed yet.


If only there was some sort of conversational agent that could be utilized to help guide people towards the plugin that they want instead of a phonebook listing.

Oh well. It’s only the year 2000. One can only dream.


Someone needs to release a plugin for plugin store seo optimization.

maybe call it “A SEO Plugin”


Better yet, we should be able to review them. Negative reviews for plugins that just cause the AI to regurgitate search results without paying attention to the query, or that return so much junk it overwhelms the context window and the AI can’t remember what we were talking about.

1 Like