(Credit: Lily Yeh; Andriy Onufriyenko / Moment via Getty Images; Blackbox / Suno / JanitorAI/ Leonard)
AI promises to democratize the artistic process, but like any other industry, a few well-known brands dominate the conversation. You've probably heard of big players like OpenAI's ChatGPT, Google's Gemini, Anthropic’s Claude, and Meta AI, but there are a number of lesser-known yet highly capable tools that may offer certain advantages over larger and more popular LLMs. Here we'll explore a few intriguing AI alternatives and how to use them.
1. Blackbox AI
(Credit: PCMag / Blackbox)Blackbox AI is designed to help software developers solve complex programming challenges and execute coding-related tasks. The idea is that with very limited prompts (or even by adding screenshots or Figma files), Blackbox AI can generate a complete full-stack application, cutting down on repetitive coding tasks and testing.
That said, it still offers a full suite of standard capabilities, like a chatbot and research assistant, and supports many popular models for free. However, a $19.99/month subscription is required to access all API models and advanced features, like the app builder and AI coding agent. You’ll also need to create an account. A premium account lets you execute “Pro” searches, which are designed to handle deeper research and multistage queries or tasks.
2. ElevenLabs
(Credit: PCMag / ElevenLabs)ElevenLabs offers a suite of tools ranging from text-to-speech to automatic dubbing into other languages, powered by a wide range of AI voices. Voices include a number of tones, pitches and accents, and are quite varied and realistic sounding. You can also generate sound effects from a text prompt, like a lion roaring on the savannah or a cat meowing in a café. Some troubling features are available, like the ability to clone a person’s voice by entering voice samples, as former President Joe Biden found out.
While a handful of examples are available without signing in, full access requires creating an account. You’ll be asked to fill in some basic information and add your intended use case for the software (PCMag's Brian Westover used it to help him create audio workouts). With a free plan, you get 10 minutes of text-to-speech or 15 minutes of conversational AI, though you can’t use advanced features like cloning or dubbing.
ElevenLabs offers five other pricing tiers, with escalating features and additional minutes of output, ranging from $5 to $1,320 per month. There’s also an Enterprise tier with custom pricing for large businesses.
3. JanitorAI
(Credit: PCMag / janitorai)JanitorAI takes the AI chatbot concept to the next level by allowing you to develop specific AI characters. These personas can be based on existing fictional characters or designed from the ground up and tailored to your specifications. The tool is currently in beta, but has received frequent updates since launching in June 2023.
Once you set up an account, click Create a Character and enter the data that will be used to generate the chatbot. It’s important to differentiate between your character’s biography, which influences how other users can find it, from personality and example dialogs, which will affect its behavior. What you enter will be used to train how your character reacts, so the more detail you add, the closer they’ll be to your vision.
When you chat with your creation or another user’s character, you’ll have to select an AI model API to power the interaction. JanitorAI's proprietary model is currently free, though premium features are in the works. A third-party API such as ChatGPT or Claude may require a purchase. OpenAI, for instance, offers 500 messages but then charges based on which model you choose.
4. Leonardo AI
(Credit: PCMag / Leonardo Interactive)Leonardo AI is a collection of AI tools designed for creating, editing, and enhancing visual art. This includes static images, like existing photos, animations, and 3D models. The platform also offers advanced features, like custom upscaling or a transparent PNG generator for background-free visuals.
To get started, you’ll need to create an account with either Google, Apple, or Microsoft credentials, or by entering your email. Once you’re signed in, select from a number of preset visual styles, then enter a text prompt to generate an image. You can also fine-tune the generation process by choosing Image Guidance, which allows you to upload reference images; the AI will attempt to mimic their styles.
Leonardo offers free access, along with three paid tiers and a custom API plan. You get 150 tokens every 24 hours on a free plan, allowing you to make a limited number of creations. Starting at $10/month, a paid plan gives you more tokens and a whole month to use them between refreshes. You also get access to more Collections, which let you organize your creations, and other advanced features. At the highest paid individual tier, you gain access to video generation and the ability to train AI models.
5. Suno
(Credit: PCMag / Suno)Suno is an AI music-creation platform built by alums from places like Meta, TikTok, and AI startup Kensho. It allows you to create songs that combine instrumental music and lyrics through text prompts. You can then save or share your music, or use premium editing tools to modify it to better suit your vision. Its ability to quickly generate complete songs in a number of styles and genres, including reasonably competent lyrics, is pretty uncanny.
Sign up with a phone number or Apple, Discord, Google, or Microsoft account. Suno offers a free version and two paid options ($10 or $30 monthly) with more access and features. The free version gets you 50 credits that renew every day (enough to make 10 songs) and lets you save up to four songs at once. A paid plan gives you 2,500 or 10,000 monthly credits and the ability to create up to 10 songs at once.
If you want to use any of your creations commercially, you must upgrade to one of the paid plans. Suno will essentially license the use of the song, which will continue as long as you remain subscribed. However, its model has run into some legal trouble, with the Record Industry Association of America suing Suno (and Udio) last year for training its AI on unlicensed music created by top artists.
An Aggregate Future?
A trend we're beginning to see from smaller operations is aggregation. If they can't compete with the heavy hitters, they'll instead offer choice, allowing you to use other companies' models all on one platform. Quora’s Poe is an excellent example of this. Its entire purpose is to let users interact with models from OpenAI, Anthropic, and Meta, all from a single, easy-to-use interface.
This could be a boon for smaller tools, like some on our list, giving them access to more advanced models or those trained with far greater resources. It can be especially useful for experimental models like JanitorAI, which can deploy powerful platforms for specific, highly customized, and narrow applications.