Every time you open the internet, someone seems to be talking about a new AI tool.
One promises to write better emails. Another creates beautiful images. And there’s always something claiming to make you ten times more productive.
But here’s what caught my attention recently.
While everyone is busy discussing ChatGPT, Claude, and Gemini, several lesser-known AI tools are doing things that seemed almost impossible just a few years ago.
Imagine an AI assistant that can operate your smartphone, edit videos, create a playable 3D game, manage your business tasks, or even help conduct scientific research.
And these aren’t just ideas from a science-fiction movie. Real tools are already exploring these possibilities.
I spent some time looking through recent AI launches and product updates. Here are 20 new AI tools in 2026 that deserve a closer look.
Some are useful for everyday work. Others are designed for creators, developers, researchers, and businesses.
You probably won’t need all of them. But I wouldn’t be surprised if at least a few make you think, “I wish I’d known about this earlier.”
1. Kino AI – Edit Videos by Simply Telling AI What You Want
Video editing can be surprisingly time-consuming.
You have to arrange clips, trim unnecessary footage, adjust scenes, and sometimes spend hours making a video look right.
Kino AI approaches video editing differently.
It’s a video editing environment that works with AI agents such as Claude Code and Codex.
Instead of making every change manually, you can ask an AI agent to help organize footage, edit sequences, or make adjustments to your video project.
Imagine you’ve recorded several clips for a YouTube video. You could ask your AI agent to find relevant footage, arrange a rough cut, and help refine the result.
You still have control over the final video, but AI can assist with the repetitive parts.
Why it’s interesting: Kino AI could be useful for YouTubers, video editors, content creators, and small businesses that regularly produce videos.
Rather than treating AI as a separate video generator, Kino brings it into the actual editing process.
Try Kino AI
2. Prisma ChatGPT Plugin – Build and Deploy Apps Without Database Headaches
Creating an application with AI is easier than ever.
But once the basic application is ready, another problem appears.
Where will you store all the information?
Every application needs somewhere to save things such as user details, orders, payments, or customer records.
Setting up those databases can be complicated, particularly for beginners.
The Prisma ChatGPT Plugin aims to simplify that process.
It connects ChatGPT with Prisma’s database and deployment tools, allowing developers to work on database setup and application deployment through natural-language instructions.
For example, imagine building an expense-tracking application. You need to store transactions, dates, expense categories, and user information.
Prisma can help handle the technical work of connecting the application to a database.
Why it’s interesting: It reduces some of the complexity involved in turning an AI-generated application into a working product.
It’s particularly useful for developers and people experimenting with AI-assisted software creation.
3. Hume Audio Decision API – AI That Understands More Than Just Words
Most AI audio tools focus on converting speech into text.
But there’s a big difference between understanding what someone says and understanding how they say it.
Think about a customer service call.
A customer might say, “That’s fine,” but their tone could suggest frustration or disappointment.
Hume’s Audio Decision API is designed to analyze audio beyond the words being spoken.
Developers can use it to ask structured questions about recordings, including characteristics of speech, vocal delivery, and sounds in the background.
For example, a customer-support company could build a system that flags calls that may need closer attention.
It could also be useful for analyzing voice assistants, reviewing recordings, or identifying important sounds.
Of course, interpreting emotions from someone’s voice is never completely reliable, so human judgment still matters.
Why it’s interesting: Hume opens up possibilities for AI applications that work with the richness of sound rather than relying only on transcripts.
4. Fireworks Nexus Desktop – Use Different AI Models in One Place
If you’ve tried several AI tools, you’ve probably noticed something.
One model might be excellent at coding. Another may be faster, while a third could be more affordable for everyday tasks.
But managing different models and subscriptions can become confusing.
Fireworks Nexus Desktop is designed to make that process easier.
It helps users connect supported AI applications and coding environments to different AI models.
For example, a developer could choose a powerful model for solving a complicated software problem and a smaller, faster model for simpler tasks.
Nexus works with tools including Codex, Claude Desktop, and OpenCode.
Instead of being limited to a single model provider, users can explore different options depending on their requirements.
Why it’s interesting: It can give developers more flexibility over AI model selection, performance, and usage costs.
For people who regularly experiment with AI coding agents, this is worth exploring.
5. Oumi FAiE – Build AI Agents That Can Keep Improving
Building an AI chatbot is one thing.
Building an AI agent that consistently performs well is something else entirely.
Imagine developing an AI assistant for customer support.
It may answer simple questions correctly, but what happens when customers ask complicated questions or provide confusing information?
You need a way to test the agent, identify weaknesses, and improve its performance.
That’s the idea behind Oumi’s Frontier AI Engineer, or FAiE.
It helps developers work through different stages of creating and improving AI agents, including evaluation, data preparation, training, and testing.
Instead of repeatedly changing prompts and hoping for better results, developers can use a more organized process.
Why it’s interesting: Oumi is exploring how AI can assist in building and improving other AI systems.
It could be valuable for businesses developing specialized AI assistants for their own operations.
Try Oumi FAiE
6. Claude for Work – Turn Meetings and Ideas Into Actual Work
Most people already know Claude as an AI assistant that can answer questions, write content, and help with coding.
But Anthropic is expanding Claude into everyday workplace activities.
Its newer workplace features include Claude Sheets, Claude Whiteboards, Claude Data, and Claude Meetings.
Imagine brainstorming with your team on a digital whiteboard.
Instead of manually organizing every idea afterward, you could ask Claude to turn the discussion into an organized summary or action plan.
Claude Sheets focuses on working with spreadsheet-style information, while Claude Meetings is designed to assist during online meetings.
These features are being introduced in stages, so availability depends on the product and account.
Why it’s interesting: Claude is becoming more than a place to ask questions.
It’s moving closer to the documents, meetings, spreadsheets, and applications people use at work.
For professionals handling multiple responsibilities, that could be particularly useful.
Try Claude
7. Adobe Creative Agent – Let AI Help With Your Design Work
Creating a professional design involves much more than generating one attractive image.
You may need to edit photographs, remove unwanted objects, adjust layouts, change backgrounds, and prepare different versions for social media.
Adobe Creative Agent aims to help with those workflows.
Adobe is bringing AI agents into its creative applications, including tools associated with Photoshop, Premiere, and Firefly.
Imagine preparing promotional material for a business.
You might want to remove a distracting background, adjust the layout, and create different versions for Instagram and a website.
Instead of switching between multiple tools and manually repeating every step, AI assistance could make the process easier.
Some Creative Agent features are available in beta, while others are being introduced across Adobe’s applications.
Why it’s interesting: Adobe is moving beyond simply generating images with AI.
It’s trying to make AI part of the complete creative workflow.
This could be particularly useful for designers, marketers, photographers, and content creators.
8. Databricks Always-on Genie – An AI Assistant That Doesn’t Wait for Questions
Most AI assistants work in a familiar way.
You ask something. They answer.
But what if an AI assistant could monitor important business information and alert you when something needed attention?
That’s the idea behind Databricks Always-on Genie.
It’s designed to handle longer business tasks, work with enterprise information, and coordinate activities that may require multiple steps.
Imagine managing a company that sells products through several stores.
Every day, you need to monitor sales, customer activity, and business performance.
Instead of repeatedly requesting reports, an always-on AI assistant could help track relevant information and highlight important developments.
It could also identify when human input is needed.
Why it’s interesting: This represents a move toward AI systems that can support ongoing business operations rather than simply responding to individual requests.
It’s mainly aimed at organizations with substantial data and complex workflows.
Try Databricks Always-on Genie
9. Paige by PagerDuty – An AI Assistant for Fixing Website Problems
Imagine running an online business.
Everything is working perfectly until your website suddenly stops responding.
Customers can’t place orders, employees start receiving complaints, and your technical team has to figure out what happened.
Finding the cause of such problems can be stressful and time-consuming.
PagerDuty has introduced Paige, an AI agent designed to assist with this type of situation.
Paige focuses on site reliability engineering, which is simply the work involved in keeping software systems running reliably.
It can help teams investigate incidents, review relevant information, and identify possible causes.
For example, if a server becomes unavailable, Paige can help narrow down the problem by examining technical information connected to the incident.
Human approval and oversight remain important when taking corrective actions.
Why it’s interesting: Paige could help technical teams respond to problems more efficiently.
This is especially valuable for businesses where even a short website outage can result in lost revenue.
10. Fred by Fireflies – Your Meeting Assistant Is Becoming an AI Teammate
Meetings often create more work than they complete.
Someone has to write the summary, identify important decisions, prepare follow-up messages, and make sure everyone knows what to do next.
Fireflies became popular by helping people record, transcribe, and summarize meetings.
Now it’s expanding that idea with Fred, an AI teammate designed to assist with broader workplace activities.
Imagine finishing a meeting with a client.
Instead of searching through your notes, you could ask Fred what was agreed, what questions remain unanswered, and what needs to happen next.
With suitable connected applications, Fred can help retrieve relevant information and support follow-up work.
Why it’s interesting: Fred aims to make meeting information useful long after a conversation ends.
For managers, freelancers, consultants, and busy professionals, that could make everyday work more organized.
11. Gemini Decisions API – AI That Can Make Thousands of Small Decisions
Here’s an interesting development from Google.
Most AI models are designed to generate answers in natural language.
Gemini Decisions API takes a more focused approach.
It helps developers build applications that make structured decisions based on information they receive.
Imagine an online store receiving thousands of product reviews.
Some customers are happy, others report problems, and a few want refunds.
Instead of manually reading every review, a business could use the Decisions API to classify information into useful categories.
One of its notable features is the ability to evaluate up to 1,000 independent questions about a shared input in a single request.
It can also work with different input types, including text, images, audio, video, and documents.
Why it’s interesting: Gemini Decisions API could be useful for businesses processing large amounts of information.
It may not look as exciting as an AI image generator, but it could become an important building block for smarter applications.
12. Nod – An AI Assistant That Can Actually Use Your Computer
What if your personal assistant could open websites, navigate applications, and handle routine tasks for you?
That’s what Nod is designed to do.
Created by n8n Labs, Nod is an open-source personal AI agent that runs on your own computer or server.
It has access to a browser and can work with connected applications such as Gmail, Slack, Notion, Google Drive, and Calendar.
Imagine you’re preparing for an important meeting.
You could ask Nod to gather relevant emails, check your calendar, and prepare a short briefing.
Or perhaps you want to cancel a subscription that you no longer use.
Nod can help work through that process.
One feature I appreciate is its emphasis on approvals. It is designed to request permission before taking certain sensitive actions.
Why it’s interesting: Nod brings AI automation closer to everyday personal and business tasks.
The software is free and open source, although AI provider costs may apply. It’s still an experimental product, so important actions should be reviewed carefully.
Try Nod
13. Wren by Seldon – Test Your Website Without Writing Complicated Scripts
Have you ever visited a website and discovered that a button doesn’t work?
Maybe the registration page freezes or the payment process fails halfway through.
These problems can frustrate customers and damage a business.
Developers normally spend considerable time testing applications to prevent such issues.
Wren by Seldon aims to make software testing easier using AI agents.
Instead of manually preparing every testing instruction, developers can describe what should happen.
For example:
“A customer should be able to register, add two products to the shopping cart, and complete payment.”
Wren can then interact with the application, explore possible paths, and identify failures.
It can also investigate unexpected situations that may not be included in ordinary tests.
Why it’s interesting: Wren could help developers discover problems earlier and spend less time writing repetitive testing scripts.
It currently offers an introductory free testing allowance.
Try Wren by Seldon
14. Spacefast – Publish Your AI-Built Application Without Complicated Setup
AI coding tools have made application development more accessible.
You can describe an idea to an AI assistant and get help building the software.
But once your application is ready, there’s another challenge.
How do you put it online?
You may need a hosting service, database, domain name, and other technical configurations.
Spacefast, developed by Automattic, aims to simplify that process.
It provides infrastructure that AI coding agents can use to publish and manage applications.
Imagine creating a small inventory management application for your business.
After building it with an AI coding assistant, Spacefast can help handle the steps needed to make it accessible online.
It supports features such as application hosting, databases, custom domains, and developer integrations.
Why it’s interesting: Spacefast helps bridge the gap between creating an application and actually using it.
For entrepreneurs and developers experimenting with AI-built software, that could save considerable time.
Try Spacefast
15. Glasshand – Let AI Operate Real iPhones and Macs
This is one of the more unusual AI tools on the list.
Glasshand allows compatible AI agents to interact with actual iPhones and Macs through an API.
And yes, we’re talking about real physical devices, not just simulations.
Normally, developers testing an iPhone application use a simulator or connect their own device.
Glasshand provides access to remotely hosted Apple devices that AI agents can operate.
The agent can open applications, tap buttons, type information, and navigate through screens.
Imagine developing a mobile banking application.
You need to check whether the registration process works correctly and whether users can navigate through different screens.
Glasshand could help automate parts of that testing on real devices.
Why it’s interesting: It opens up possibilities for mobile automation and real-device software testing.
It’s mainly useful for developers and technical teams, rather than ordinary smartphone users.
Try Glasshand
16. Wanaka Studio – Create Your Own 3D Game by Describing an Idea
Have you ever dreamed of building your own video game?
Perhaps you imagined a racing game, a fantasy adventure, or a virtual world where players could explore different locations.
The problem is that game development normally requires programming knowledge and experience with complicated software.
Wanaka Studio is trying to make that easier.
It’s an AI-powered platform that can turn written ideas into playable 3D games.
You describe the game you want, and its AI agent helps create the world and gameplay.
For example, you could write:
“Create a racing game where players drive through a futuristic city at night.”
Wanaka can help turn that description into an interactive game.
You can then use its visual editing features to improve the result.
Why it’s interesting: Wanaka makes game creation more approachable for students, hobbyists, teachers, and people who have creative ideas but limited coding experience.
The exciting part is that it creates playable experiences rather than simply producing videos that look like games.
Try Wanaka Studio
17. Beacon.sh – Give Your AI Coding Assistant a Long-Term Memory
Anyone who regularly works with AI coding assistants knows how frustrating repeated explanations can be.
You spend time explaining your project, correcting mistakes, and describing how certain features should work.
Then you start another session, and the AI doesn’t always have access to what it learned before.
Beacon.sh is designed to address that problem.
It’s an open-source memory system for AI coding agents.
It captures useful information from previous sessions and turns important lessons into reusable knowledge.
It supports tools such as Claude Code, Cursor, Codex, and OpenCode.
Imagine spending several weeks building a business application.
Your AI assistant gradually learns how your project is structured and how you prefer certain tasks to be completed.
Beacon can help preserve those useful discoveries so future sessions can build on them.
Why it’s interesting: It could make long-term AI-assisted coding projects more consistent and reduce the need to explain the same things repeatedly.
Try Beacon.sh
18. ClawsGO Science – An AI Agent That Can Help Conduct Research
Scientific research isn’t just about finding information online.
Researchers must review published studies, analyze data, perform calculations, prepare charts, and document their findings.
ClawsGO Science aims to assist with several of these activities.
It’s an AI research agent that can work through different stages of a research project.
For example, a researcher studying climate change could ask it to find relevant scientific publications, examine datasets, and help prepare a research report.
ClawsGO can also execute code, generate figures, and maintain records of its activities.
One particularly interesting feature is its ability to continue working in a cloud environment while the researcher is away from their computer.
It also keeps an activity history so users can review what the AI has done.
Why it’s interesting: ClawsGO could be useful for university students, scientists, and researchers handling complicated projects.
However, AI-generated findings should never be accepted without verification. Researchers must still check sources, methods, and conclusions.
Try ClawsGO Science
19. Mika – An AI Agent That Can Control Your Android Phone
We use smartphones for almost everything.
Shopping, messaging, banking, social media, entertainment, and work are all handled through mobile applications.
But imagine giving your phone a task in ordinary language and letting an AI agent navigate the required applications.
Mika is exploring exactly that.
It’s an open-source AI agent for Android that can interact with mobile apps and websites.
It can tap buttons, scroll, type, and move between different screens.
For example, you could ask it to search for products within a certain budget or gather information from a mobile application.
One interesting feature is its ability to perform certain interactions on a separate background screen, depending on the device and setup.
That means you may be able to continue using your phone while the agent works.
Why it’s interesting: Mika shows how AI assistants could eventually become more deeply integrated into our smartphones.
It’s still evolving and requires technical configuration, but the possibilities are fascinating.
Try Mika on GitHub
20. Phonon by Gradium – Generate AI Voices Without an Internet Connection
Most AI voice generators depend on the internet.
You send text to a cloud service, and the service returns spoken audio.
But what happens when your internet connection is slow or unavailable?
Phonon by Gradium offers an interesting alternative.
It’s a family of lightweight, open-weight text-to-speech models designed to run directly on devices.
The smallest English model has approximately 31 million parameters and takes up around 64 MB of storage.
Once installed, it can generate speech locally without constantly sending requests to an online server.
Imagine an educational application that reads lessons aloud even when a student has no internet connection.
Or a personal voice assistant that can respond without sending every request to a remote service.
Phonon could help make those applications possible.
Why it’s interesting: It could be useful for offline voice assistants, accessibility tools, mobile applications, and privacy-conscious software.
The English model is available under the Apache 2.0 licence, making it particularly attractive for developers.
Which of These 20 New AI Tools Should You Try First?
With so many AI tools appearing, choosing where to begin can feel overwhelming.
My suggestion is simple: don’t choose a tool just because it’s new. Choose one that solves a problem you actually have.
If you’re interested in personal productivity, take a look at Nod, Fred by Fireflies, and Claude’s workplace features.
If you’re a content creator or designer, Kino AI, Adobe Creative Agent, and Wanaka Studio could be interesting starting points.
For software developers, Prisma, Wren, Spacefast, Beacon.sh, Glasshand, Fireworks Nexus Desktop, Oumi FAiE, and Paige address different technical challenges.
If you’re interested in business intelligence and automation, Gemini Decisions API, Databricks Always-on Genie, and Hume Audio Decision API offer possibilities worth exploring.
For research and mobile technology, ClawsGO Science, Mika, and Phonon stand out.
The important thing is to explore how these tools fit into your work rather than trying to use every new AI product you discover.
Are These New AI Tools Free?
Some are free or open source, while others offer paid plans, limited free access, or developer-focused pricing.
For example, Nod and Mika are open-source projects, and Beacon.sh offers a free starting option.
Wren has an introductory free testing allowance, while specialized services such as Glasshand may involve usage charges.
It’s also important to remember that free AI software may still require a paid AI model or API.
Always check the latest pricing and availability on the official website before getting started.
Final Thoughts: The Next AI Revolution Is Already Taking Shape
After exploring these 20 tools, one thing becomes clear.
AI is moving far beyond the traditional chatbot.
We’re seeing AI agents that can operate computers, interact with smartphones, edit videos, build games, manage software systems, analyze recordings, and assist with scientific research.
And what makes this particularly interesting is that AI is gradually becoming part of the tools we already use.
A few years ago, using AI usually meant opening a chatbot and asking a question.
Now, AI can help carry out actual tasks inside applications and digital environments.
Of course, not every new AI product will become successful. Some of these tools are still in their early stages, and others require technical knowledge or careful human supervision.
But that’s part of what makes this period exciting.
We are watching AI evolve from something that mainly gives us information into something that can increasingly help us act on it.
The future of AI may not be about asking better questions. It may be about getting more meaningful work done.
And these 20 new AI tools offer a fascinating glimpse of that future.
Which of these AI tools impressed you the most? Have you tried any of them already? Share your experience in the comments.
Follow Our WhatsApp Channel or Subscribe us or at Knowlab
