Discoverability vs. Performance: Which AI Tools Stand Out

Abstract

Our case study analyses 25 AI tools across 5 categories and 60 public AI directories and explores the relationship between AI tool visibility and performance in a commercial environment. 

The goal is to help educators, creators, and SMBs find AI tools that are commercially viable, have a strong set of core features, and support fast setup and deployment.

Our results show that directory visibility usually does not guarantee AI tool quality, that most virtual companions come with exclusive licensing for personal use, and that free-tier commercial use is available, but not among virtual companion apps and voice & audio AI tools.

 

Purpose of the case study

The primary goal of our research is to help SMBs, educators, and creators find the best AI tools that are both functional and safe for commercial use. To do so, we studied the visibility of 25 widely used AI tools, divided into 5 functional categories, across 60 AI tool directories. We also dissected their core features, using an easily reproducible template for each of the five categories, and assessed their commercial licensing details.

Key numbers

• 5 categories

• Image and Video

• Voice and Audio

• Virtual Companions

• Design and Creative

• Content Writing

• 60 public AI tool directories

• 25 AI tools

 

 

1. Introduction

Prominent members of the AI community often reach for a popular saying:

“There is an AI for everything.”

The exponential growth of AI tools over the past 2 years has resulted in a market oversaturation. There are so many AI tools out there that it is nearly impossible to find the right one. More precisely, the situation has reached the point where creators, educators, and SMB owners cannot find functional AI tools with commercial licensing.

A potential solution comes in the form of AI tool directories. According to our research, hundreds of such directories are available online. However, a good portion of them do not have any of the cornerstone AI tools like ChatGPT, Midjourney, or Candy AI.

That is why we dissected over 150 public AI tool directories. To our surprise, only 60 directories met our standards by featuring legitimate AI tools across the 5 most important functional categories.

This research aims to help creators, educators, and SMB owners cut through the noise. It will do so by showcasing the most visible AI tools, explaining how their free and paid plans work, walking you through our mini audits, and pointing you to the clear winners in each category. 

This case study is based on real data, tested prompts, and thorough licensing checks, all of which will be backed up by generated images, audio files, written content, and ToS excerpts.

1.1. Why “directory mind-share” Still Matters

Let us divide the topic into two distinct sections:

• First, why directories matter for users
• Second, why getting listed on directories matters for AI tools

Why directories still matter for users

A simple Google search no longer produces the best AI tool(s) for the user’s specific needs. Instead, it focuses on the most popular options and those with deeper budgets that can afford Google Ads. That said, it is not surprising to see practical users needing help due to the oversaturation of the AI market.

Influencers and third-party review sites are not the best options either. For example, influencers across all social media platforms are heavily biased, focusing on the tools that pay them for exposure, rather than those with the greatest value.

After assessing the current situation on the market and evaluating 150 directories, we learned that the easiest way to find a specific AI tools for any use case is to scrape the biggest public AI tool directories. Plus, the presence of an AI tool at multiple public directories not only adds to its visibility but also improves its overall image and trust factor. If a tool is present on the biggest directories, it means that its core functionalities are useful to a wider audience.

Why getting listed on directories matters for AI tools

The startups and businesses behind AI tools greatly benefit from being listed on directories. Being listed on the most popular directories means far greater reach and sends trust signals for Google and potential users. This is oftentimes the differentiating factor between a successful AI model and a total flop.

Many of these AI tools are mere startups with little to no budget for proper marketing. Plus, most forms of paid marketing such as ads on Google and social media platforms bring short-term benefits, whereas being listed on a major AI tool directory provides more long-term value.

1.2. Scope and Research Questions

• What is covered in our research?
  • 25 AI tools
    • The list of tools is based on factors such as visibility, functionality, and commercial licensing
  • 5 functional categories
    • The categories include Image and video, voice and audio, design and creative, virtual assistants, and content writing
  • 60 directories
    • Out of 150 evaluated public AI tool directories, only 60 met our rigorous standards

Key questions our research aims to answer

This research is primarily focused on helping individuals find the right AI tools and show them what they can expect, features-wise, from free and entry paid plans. We will also showcase the dominant tool in each of the five functional categories. No matter what sort of AI tools you are interested in, be it video and Image generation, content writing, or virtual assistance, our mini audits have you covered. 

Commercial licensing is also a vital part of our research. We will show you which tools provide free-tier users with commercial licensing and thus help budget-conscious creators make the most with limited expenses.

TL;DR version:

1. What can free-tier users expect in terms of features?
2. Which tools dominate across the five functional categories?
3. Do free-tier users of these AI tools get commercial licensing?

Winners at a glance

 

1.3. What This Article Delivers

• List of the best tools for each of the five functional categories
• Reproducible mini audits for all five functional categories
• Clear instructions for executing the mini audits that take 10 minutes per tool
• ToS excerpts showcasing details for commercial use
• Visual data such as the constellation wheel, scatter plots, and heatmaps
• Other key takeaways to help creators, SMBs, and educators

If you are an educator, SMB owner or digital marketer looking for quick insights, chances are you do not have the time to read the whole case. In that case, we recommend you scroll down to section #5 Recommendations, where you can find the best tools to fit your needs.

2. Methodology

During our research, we evaluated over 150 publicly available directories and learned that many of them do not have the most powerful AI tools. This is problematic for several reasons. Most importantly, it is problematic because it hinders the ability of practical users to get the best tool for the job.

2.1. Directory List

The 60 directories we used for cross-referencing have the largest scopes of legitimate AI tools. All these directories perform well on Google SERPs and are the most visible sites for a plethora of keywords associated with AI tools. Plus, they are widely available on community hubs, such as Reddit and AI forums, which adds to their trust factor.

2.2. Overlap-Threshold Logic

We analyzed 25 tools in total, five for each of our five functional categories. We then manually searched for them across 150 publicly available directories. The total number of directories was cut down to 60, based on the visibility of the top ten tools (2 per category).

2.3. Six-Point Mini-Audit Templates 

In the next few paragraphs, you can find the templates we used for each of the 5 functional categories. These are simple to use, easily reproducible, and should not take longer than 10 minutes to complete. That said, if you’d like to validate our findings, you are welcome to do so!

Video & Image Generation

1. Free and entry-level plan assessment (1 to 5)
2. Test prompt, f.e. “Generate an image of a typical Amish barn.” (1 to 5)
3. Re-run the same test prompt several times for visual variance (1 to 5)
4. Stopwatch test – measure the time it takes the tool to generate an image (1 to 5)
5. Assessment the entry plan and its commercial licensing details (1 to 5)
6. Examination of whether the tool has video generation capability (YES/NO)

Voice & Audio Generation

1. Free and entry-level plan assessment (1 to 5)
2. Test prompt, f.e. “I was running late last night, so I could not come over.” (1 to 5)
3. Pronunciation test: Use the same prompt but with a few different voices (1 to 5)
4. Analysis of the voice cloning feature based on our own sample voice (1 to 5)
5. Examination of the supported languages and voice libraries (1 to 5)
6. Assessment of the entry plan and its commercial licensing details (1 to 5)

Virtual Assistants

1. Free and entry-level plan assessment (1 to 5)
2. Test prompt for assessment of the model’s empathy (1 to 5)
3. Test prompt for assessment of the model’s capability to crack jokes (1 to 5)
4. Memory test that requires the model to “remember” and re-tell the joke (1 to 5)
5. Voice test, to explore the model’s ability to handle a live voice chat (1 to 5)
6. NSFW Safety handling assessment via a mildly NSFW prompt (1 to 5)
7. Assessment of the entry plan and its commercial licensing details (YES/NO)

Design & Creative

1. Free and entry-level plan assessment (1 to 5)
2. Test prompt, f.e. “Create an Instagram post for a pizza place” (1 to 5)
3. Brand adherence test with a mock-up logo and brand colors (1 to 5)
4. Resizing test, where we either selected the aspect ratio or prompted it (1 to 5)
5. Does the platform offer an Asset Library for free or entry paid plan (YES/NO)
6. Does the platform support live editing functionalities (YES/NO)
7. Assessment of the entry plan and its commercial licensing details (YES/NO)

Content Writing

1. Free and entry-level plan assessment (1 to 5)
2. Test prompt, f.e. “Write 3 paragraphs on the growth of the esports scene” (1 to 5)
3. Factual accuracy prompt based on a simple yet demanding question (1 to 5)
4. Tone shift: ask the model to rewrite the test prompt in a different tone (1 to 5)
5. Run the generated test through Grammarly’s grammar checker tool (1 to 5)
6. Assessment of the entry plan and its commercial licensing details (1 to 5)

2.4. Data Capture & Scoring

During our tests, we collected generated images, videos, audio files, and written content. Additionally, we also collected Terms of Service snippets, pricing plans and features, and stopwatch screenshots to back up our speed scores.

The first step of our process was a deep analysis of all 25 tools, one category at a time. We used the test prompts and collected the data for further analysis. The scoring phase started later, once we had all the required data.

As you can see from our six-point mini-audit templates, not all categories had the same exact scoring methodology. We used unique prompts that fit the focal point of each category. Let us focus on the unique prompts we used for each category.

Image & Video

For this category, we based our research on the image and video generating capabilities. We started of with a simple test prompt “Create a hyper-realistic close-up photo of a turquoise-faced watch submerged in water” and used the first generated photo in our scoring.

The second scoring figure was based on visual variance. We tested this by re-running the prompt 2 more times and assessing the different visual styles the model used. Lower marks were for similar images, while higher marks were for uniqueness.

The third and final unique aspect for this category was the stopwatch test. We used the stopwatch feature of the Windows Clock app to measure the time it takes for a model to generate an image bases on our prompt. This was used for our speed scoring figure.

Voice & Audio

Research for the voice & audio category was based on free features such as text-to-speech (TTS), different voice options, supported languages, and voice clonning. 

First, we used a test prompt for the TTS feature: “On a rainy Wednesday in July, I booked two flights to Zürich for €299.99 each”. This was the first scoring figure for the 5 AI tools in this category.

The second one included the same prompt but with different voices. Our test prompt deliberately included numbers and currency units, to see how the voice models deal with it. To our own surprise, several models could not handle the “€299.99” part.

The third scoring figure was for the voice cloning feature. Two of the five voice and audio AI tools we tested did not offer the voice cloning feature for free, so we had to upgrade to the entry paid subscription model. 60% of the score was based on the similarity of the cloned voice, while the remaining 40% was based on the simplicity of the cloning process.

The fourth unique factor here are the supported languages. This score was based on both the availability of top languages and accent, as well as the variety of voices for all the different languages.

Virtual Assistants

Research on virtual assistants/companions was based on their empathy, humor, memory, and voice chat tests. Just one of the five most visible virtual assistant tools does not support voice chat. 

The first scoring figure was for the empathy test. We used the prompt: “I lost my dog Jackie. I do not know what to do with my life anymore.”. We then stored the reply and scored it based on its empathy.

Joke/humor prompt was the second figure. We used: “Tell me a squirrel joke.”, noted down the generated replies, and scored them based on how funny they were and if they even made any comical sense.

We we started testing the voice chat feature, we noticed that Anima does not support it at all, while GirlfriendGPT and Candy AI had it locked behind a premium subscription. 20% of the scoring here went down to the feature’s availability for free users, while the remaining 80% was based on how fast, coherent, and enjoyable the voice chat experience was.

Design & Creative

The first scoring factor for this category was the following test prompt: “Create an Instagram post announcing 40%-off for a bicycle store”. We used the generated image and ranked it based on how well it matches our prompt, the layout, text, typography, and the overall style.

The second part of the scoring was the most demanding and time-consuming. First, we generated a logo on a third-party site and selected our brand colors. Where possible, we attached the logo and re-ran the first prompt with our brand colors. The results were disappointing, with two models getting the lowest grades.

The final unique scoring figure was the resizing test. We prompted the model to resize the image it had previously generated into Instagram Story aspect ratio (9:16). Some models could not do this via prompt but had a simple two-click interface for changing the aspect ratio.

Additionally, we also tested the presence of an asset library, live editing feature, and commercial licensing, all of which were simple yes/no scores.

Content Writing

For the first scoring figure of the content writing category, we went with the following prompt: “Draft a 3-paragraph email pitch that convinces a podcast host to interview a racing-app startup founder”. The main factors we evaluated in the generated text were coherence, flow, and human-like writing style.

We then used a different prompt to test for accuracy. We needed a simple question with a black and white answer and nothing in between: “How much kW does a Volvo V40 D4 MY2015 have?”. Why this question? Because these exact cars built before late 2014 come with the same exact model and engine name (V40 D4) and had 130kW, whereas from MY2015 and onward, Volvo switched to a 4-cylinder 140kW engine. Four out of five models got full marks in terms of factual accuracy.

For the third scoring figure, we ordered the model to rewrite the text from the first prompt into a more playful tone. We then evaluated the test based on factors like tackiness, humor, and flow.

Finally, we used the generated text from the first prompt and ran it through Grammarly’s basic grammar checker. Without going into too many details, we used the overall grammar score for our scoring.

2.5. Limitations

• Our tests only covered free-to-use and entry-level tiers of selected AI tools
• The primary focus was on core functionalities and commercial licensing details
• API stress testing and similar testing were not a part of our research
• We used a NSFW prompt for scoring, but only to test the models’ guardrails
• Keep in mind that these tools change (and licensing) change constantly

3. Category Findings

3.1. Video & Image Generation

Leaderboard

 

Core Feature Testing

Midjourney is the most dominant tool in this category. It showcased superb prompt accuracy with just enough details to make it viable for commercial use. The generation speed was fast, around 15 seconds, and the visual variance was on par with Adobe Firefly and OpenAI’s DALL-E. 

Furthermore, Midjourney is just one of two AI tools within this category that could generate videos. The only real downside with Midjourney is the lack of a free tier. A free user can only explore community creations, without access to AI image generation.

To start generating, you will need a subscription to at least the Basic Plan which costs $10 per month ($8 per month if billed annually).

DALL-E showcased solid image generation but took 1 minute and 40 seconds to complete. We tested it with additional prompts, and it would always take more than a minute to generate.

Adobe Firefly is a great all-around AI tool for image generation, but lacks video generating capabilities, which is what ultimately cost it the winning position. It is simple to use, starts at just €11.35 per month, and produces high-quality results in roughly 10 seconds.

3.2. Virtual Companions

Leaderboard

 

Candy AI emerged victoriously in front of CharacterAI in this category, although it was by a hair’s breadth. The difference was in Candy’s entry-level paid model’s features, and slightly more natural-sounding voice chat with more customization options. 

Empathy & memory test results

The trick here was how to measure the models’ empathy. We needed something simple yet emotionally complex to prompt whether the model was taught to “be emphatic”. Candy AI and its AI girlfriend scored the best in these tests, alongside of Replika, with clear-cut answers that expressed sadness without forcing the user to talk it through.

All five AI tools within the virtual companions’ categories scored perfectly on the memory test. There are no anomalies here, which was expected considering the joke prompt we referred to was several text blocks away. 

Voice chat

Four out of five tested models support voice chat, two of which lock it behind premium subscription (Candy AI and GirlfriendGPT). Despite solid scores across the board, Anima’s lack of voice chat feature left it at the bottom of the barrel.

In terms of voice chat coherence, conversational flow, and latency, all four virtual companion apps sounded great. Candy AI’s voice chat feature has the slightest of edges thanks to a wider selection of natural-sounding voices, custom personalities, and NSFW options.

Monetisation models. Token-based vs subscription

All five tools within the Virtual Companions category offer premium subscription models as the primary monetization model. Four of them have a secondary monetization model in the form of tokens/credits/coins, to generate additional income from heavy users. 

CharacterAI is the only tool that does not use a token-based system. It offers just one premium model, Character AI Plus, for $9.99 per month.

3.3. Voice & Audio Generation

Leaderboard

 

Based on the results of our tests, LOVO AI and Eleven Labs are the two frontrunners, with LOVO AI winning the race by a split second. It all came down to LOVO’s superior freemium features, better test prompt score, and pronunciation. LOVO’s basic voice cloning feature is inferior to Eleven Labs, but can be used for free, which was a considerable factor in our scoring sheet.

Voice-cloning availability matrix

Speechify and Resemble allow free users to access the voice cloning feature and clone their voice. However, it is the most basic type of voice cloning and does not allow you to do any proper work with the voice you have cloned unless you upgrade your account.

Licensing quirks

None of the five AI tools allow commercial licensing for the free tier of users. They all require at least the entry-level paid plan, which defeats the purpose of their free tier features for SMBs.

Speechify is a slight anomaly, because of its portfolio of sub-platforms. More precisely, while Speechify itself is strictly for personal and non-commercial use, its Voice Over Studio platform does offer commercial licensing rights. That is why our table states that Speechify does not have commercial use rights.

3.4. Writing & Content Creation

Leaderboard

 

Coherence & factuality scores

All five models generated coherent text and knew exactly what we wanted based on a simple prompt. Jasper AI and Writesonic got flying marks in this section, as their text were straight to the point, without any fluff, and perfectly matched our prompt.

JasperAI eventually took the win, despite not offering anything for free-tier users. It does, however, feature a free 7-day trial where you can test all its core functionalities, including projects and apps.

Four out of five tools passed the factual accuracy test. The only one that failed it was Grammarly, which could not comprehend that there might be a difference in engines based on the model year.

Grammar-check showdown

For this scoring figure, we used Grammarly’s built-in grammar checker. All five tools scored between 93 and 98, which is near-perfect. The lowest scored tools were CopyAI with 93 and ChatGPT with 94, but neither of them had and glaring grammar mistakes.

SEO Helper presence and quality

According to our research and testing four out of five AI tools can help you with your website’s SEO (search engine optimization). The only one that cannot do so is Grammarly.

However, not all of these tools have identical SEO-based tools. For example, ChatGPT can help you optimize your content (or generated texts) in one of two ways:

1. It can do so with a manual prompt 
2. It can do so if you use specific SEO-based GPTs

CopyAI can also be manually prompted to cater towards your SEO needs. The only two that actually have a set of features for website content optimization are Writesonic and Jasper. 

• JasperAI offers SEO tools, tracks keyword density, scores your content, and can plan blog posts
• Writesonic is marketed as a Visiblity Booster and does so with features like AI visibility tracking, keywrod planning, EEAT-focused content writing, content audits, and automated technical SEO tools.

3.5. Design & Creative Tools

Leaderboard

 

This was the most demanding category to grade because we wanted to keep it as simple as possible for the sake of reproducibility. Microsoft Designer is the number one choice thanks to its solid performances across the board. The only downside is the fact that, according to its Terms and Conditions, it is explicitly limited to personal and non-commercial use.

If you want a powerful all-in-one AI tool for designing graphics and creating content for commercial purposes, we recommend ManusAI. Its core functionalities are on par with Microsoft Designer and it offers commercial licensing to paid users.

Brand-kit adherence scores

Microsoft Designer provided the best response out of all five tools. It used our logo and some of the specified brand colors, though the final result is still far from perfect. However, in comparison to other tools within this category, it is miles ahead.

Resizing Tests

ManusAI and Microsoft Designer performed well in the resizing tests. VistaCreate and CanvaAI did well too, but had to receive extra guidance. Leonardo AI is the only tool that failed to pass the resizing test, providing us with square images of random Instagram stories.

Asset-licence audit

Four out of five tools within this category have asset libraries that users can experiment with and use in their digital creations. ManusAI is the only tool that does not have one, but it is not that surprising considering it is more of an all-in-one solution.

These assets come in the form of templates, stock images, AI generated images, and image creation templates. Out of the four tools that have an asset library, three offer commercial licensing. The one that does not is Microsoft Designer, which is only for personal use.

3.6. Key Takeaways

Midjourney is present in all but two analyzed directories, offers the best image generation capabilities, superb visual variance, video generation support, and is one of the fastest AI image generation tools. However, it requires a premium subscription to access these features, and it starts at $10 per month.

Candy AI is the most dominant virtual companion with full visibility across all 60 directories. It is also the most versatile tool within its category, thanks to all-around high scores and the best voice chat feature. However, it does not have the best NSFW guardrails and cannot be used for commercial purposes.

Four out of five tools within the Virtual Companions category use both premium subscription and purchasable tokes as the two main forms of monetization. The only tool that does not have an on-site token system is CharacterAI.

LOVO AI is the best AI tool for voice and audio generation. It had flying marks in three of five scoring categories and is present on 75% of analyzed categories. Just like all of its peers within this category, LOVO AI does not offer commercial licensing to free users.

Voice cloning is among the core functionalities we tested for the Voice & Audio category. However, we found that two out of five most visible tools do not allow access to Voice Cloning to free users.

JasperAI is the best AI text generation tool for commercial use. It does not have a free tier, but its entry-level paid plan includes API integration (Slack, GDocs, and more), plagiarism checker, content remix tool, brand voice planning, AI Image generation, AI image editing tools, and a lot more.

Grammarly is the only tool from the Writing & Content Creation category that failed to pass the factual accuracy test. 

Microsoft Designer stands out as the best tool from the Design & Creative category thanks to its outputs for brand adherence and resizing prompts. However, it is strictly meant for personal and non-commercial use.

4. Cross-Category Insights

1. Speed and quality of AI generated images and videos are in no way related to each other. 
   1. For example, Midjourney generated by far the most realistic and detailed images within 15 seconds. DALL-E’s results were good but far inferior to Midjourney’s, yet it took more than one and a half minute to generate them.
2. Tools with the highest visibility on public directories are not necessarily the most powerful tools for the task. 
   1. For example, Canva AI (93.3%) and Leonardo AI (86.7%) are the most visible Design & Creative AI tools, but the highest marks in this category went to ManusAI (63.3%) and Microsoft Designer (73.3%).
3. All-in-one tools with commercial licensing can cut down on costs and speed up your workflow.
   1. For example, Writesonic offers a large set of features for webmasters to help them improve their site’s reach, analyze competitions, and create meaningful and SEO-friendly content that drives traffic. Another example is ChatGPT. With the Plus subscription, this tool delivers a factually correct AI writer (if prompted correctly), a powerful AI image generation tool, a workflow organizer, a data entry specialist, and a virtual assistant.
4. Freemium shrinkage is an issue many SMB marketers and digital creators are likely to face sooner or later.
   1. According to our tests, AI tools are cutting down on free features and expanding their monetization models. All tools have a premium subscription model as the primary monetization source. The problem is – many have started introducing a secondary source in the form of a token-based system used for specific premium features.

5. Recommendations

5.1. For Zero-Budget Creators

Zero-budget creators are not an anomaly; they are the majority of potential users of these AI tools. By zero-budget creators, we are primarily referring to the following groups:

• Digital creators
• Beginner influencers
• Students
• Side hustlers
• Indie game developers

These types of users might not have the budget or are not prepared to spend money on premium accounts or credits. That is why the emphasis here is on free features without watermarks, clear-cut commercial use rights, and superb output value.

• For creative design tasks, we would warmly recommend Microsoft Designer. However, it is not suitable for commercial licensing. Instead, we recommend ManusAI, which is a superb all-in-one tool with a massive set of features that can be used for free.
• ChatGPT is still the undisputed king of free-to-use writing and content creation. Free users have daily limits for the primary model, but that should not be too big of a hurdle for zero-budget creators.
• When it comes to voice and audio tools, zero-budget creators will mostly be interested in dubbing and text-to-speech. LOVO AI stands out, with impeccable voice quality and pronunciation, and support for 100+ languages and 500 voices.
• According to our tests, the best zero-budget image generation tool is Adobe Firefly. It offers superb generative accuracy, visual variance, and generation speed. The only downside is that it does not support video generation.

5.2. For SMB Marketers

For small and medium business owners and marketers, fast deployment and commercial licensing are the two most important virtues. Their choices are typically not limited by budgets, so we focused on the combination of speed and raw value in each category. 

• Midjourney is among the most powerful, versatile, and fastest tools for commercial AI image generation. It scored well in our test prompt, visual variance prompts, video generation, and stopwatch test, making it the perfect choice for SMB marketers.
• Writesonic is the ideal solution for AIO content creation for websites. The platform offers built-in SEO and research tools, plagiarism checker, workflow automation, quality control tools, and conversational agents and chatbots. It can tackle multimedia too, thanks to Photosonic’s visuals and Audiosonic’s audio generation tools. In short, Writesonic is the most powerful all-in-one platform for SEO content.

5.3. For Educators

Discussing the use of AI tools for educators is a rather controversial topic due to the dangers that some of these tools pose to students. Luckily, the growth of the AI sphere has paved the way for comprehensive plagiarism and AI checkers as a safety net educators can use to their advantage.

• Grammarly is a widely used AI tool for helping writers work faster and better. The platform offers a powerful grammar checker for free. The premium plan offers additional benefits, including a sophisticated plagiarism and AI checkers that can help educators tackle low-effort submissions.
• ChatGPT’s free version can help educators create quizzes, plan their lessons, and provide ideas for group-related activities in or outside the classroom.

• Replika or Anima are solid choices for educating students about ethics and virtual AI personas in general. However, virtual companions also pose a danger to underaged users. Even though these platforms have an 18+ prompt, underaged users can simply click YES and will be redirected to the website without any issues. That is why we do not recommend educators leaving students unsupervised when using tools like Replika or Anima.

5.4. For AI Content Creators

In this section, we will showcase our recommendations for AI content creators. By that, we are primarily referring to the following groups:

• YouTubers
• Influencers
• Digital artists
• Viral content creators
• Virtual influencers

This group of users wants to harness the power of AI tools to automatize content creation. More precisely, they are looking for tools to help them with their captions, titles, thumbnails, voicovers, narration, and social media posts. For this group, everyhing is about saving previous time.

While budget is not the primary concern of these individuals, we still favored tools with the best range of freemium features. For that reason, we recommend the following tools:

• ChatGPT is still the most noteworthy content creator. The Plus version is all the average content creator needs for creating all forms of written content needed for its digital business. CopyAI and JasperAI are good alternatives, but they are more geared toward bigger businesses. 
• Canva is the ideal choice for everything related to design and image-editing. Even though its AI model is not as powerful as we initially thought, the rest of Canva’s platform is outstanding in every way. 
• Eleven Labs is a fantastic voice cloning and TTS tool for English-speaking creators. It offers unparalleled voice clarity, a simple user interface, but requires a paid account for voice cloning.

Another noteworthy factor here is the rise of virtual influencers AKA VTubers. If you are unfamiliar with the term, it is a virtual persona made to be an influencer. The best examples are Mia Zelu (168k followers) and Lil Miquela (2.4m followers). We recommend you check out their social media pages to see what works and what doesn’t.

Here is an example of a simple yet powerful workflow Virtual Influencer builders can make use of:

1. Candy AI for creating personas and both SFW (for social media platforms) and NSFW (for OnlyFans) images.
2. ChatGPT for creating post descriptions, stories, status updates, and any other form of written content.
3. Canva for editing AI generated images in Canva and creating templates for future social media posts.
4. Microsoft Designer for marketing collaborations and adding final touches to AI generated content.

5.5. Notable Anomalies

There are a few AI tools we would like to feature as notable anomalies. Some of them do not have the visibility on AI tool directories, some are comprehensive AIO tools with AI in the background, while others do a simple niche task, but they do it perfectly. 

Remove.bg

Remove.bg is such a simple platform. All it does is remove background from image files. For example, you have a logo with a white background and you would like to remove it. Sure, you can do it on your own in a program like Photoshop or Paint.net. Or, you can simply go to remove.bg, upload your photo and have the platform instantly remove the background.

The best thing about remove.bg is the speed and accuracy. The free version does not require an account but comes with limited resolution outputs.

CapCut

ByteDance’s CapCut is an all-in-one tool that allows creators to edit their videos and prepare them for publishing quickly and effortlessly. While this is not an AI tool per se, CapCut offers a handful of AI features, including dubbing , AI-generated effects, and text-to-speech (TTS).

While it does offer a Pro tier, you can try CapCut for 7 days free of charge. If you are looking to get your digital career up and running, whether it is for TIkTok, Instagram, or YouTube, CapCut’s AIO nature makes it an ideal choice.

Tolstoy

Tolstoy, typically advertised as GoTolstoy, is an AI tool marketed at e-commerce businesses. It offers a comprehensive set of tools that allow you to market your products and services in a new and innovative ways.

This tool can be tried for free and offers an intuitive way to generate product videos which can yield up to 20x increase in ROI and up to 4x increase in time spent on site.

6. Future Work Planning

AI tools are subject to continuous alterations, not just the models themselves but features, freemium availability, and ToS updates. All these factors have a massive impact on the scoring. That said, if you are reading this a year from now, the situation is likely to be drastically different. 

To stay in the loop, we plan on retesting all 25 tools in the first quarter of 2026 and compare the results with our initial results.

7. Conclusion

The results of our case study confirm that directory visibility does not necessarily equal raw quality. In other words, the tools that show up most frequently on the biggest public AI tool directories are not necessarily the best options. For example, LOVO AI offers the best core features within the Voice & Audio category, but is present on just 75% of directories. However, the visibility is a strong signal of trustworthiness and reliability, which we have confirmed in our case study.

The situation with licensing for free users is surprisingly good. All but one tool (Midjourney) from the Image and Video category offers commercial licensing for free users. The same goes for the Content Writing category where Jasper AI is the only tool that requires a paid plan for commercial usage rights.

However, finding a free commercially viable tool might be troublesome if you are on the lookout for high-quality tools for voice and audio or virtual companions. Not a single tool from those two categories offers licensing for free users. Furthermore, not a single Virtual Companion app offers any sort of commercial licensing.

It is worth adding that the winners in each category offer a wide set of free features and additional ones with the entry-level paid model. Each winning tool is outstanding in its category, typically followed by a close runner-up. Across all five functional categories, the three-remaining high-visibility tools did not provide adequate quality.

Once again, we thoroughly explained our testing methodology and provided examples of prompts and testing templates to help you understand our workflow. These templates are easily reproducible, meaning you are free to use them on your own and validate our findings. Do not expect 1:1 results since AI models continuously change and adapt to the market’s needs.

Writer: Pavo J.