ai.com
AI.com is a new personal AI service that gives each user an agent with its own virtual computer, letting it use apps and complete online tasks with less step-by-step help.
What is AI.com?
AI.com is an AI agent platform that launched publicly in February 2026.
The product is still marked as beta, so its features and rules may change quickly.
It is run by Mentat Forge, Inc. in the United States, with a related company called Mentat Forge Research Pte. Ltd. handling some users outside the country.
Kris Marszalek, the co-founder and CEO of Crypto.com, leads both AI.com and Crypto.com.
The site says its long-term goal is to build a decentralized network of agents that learn, improve, and perform useful work.
That is a much larger promise than the product available today.
What can an AI.com agent do?
Each agent receives a cloud-based computer that can work with websites, applications, and connected accounts.
A user can ask it to send messages, manage a calendar, summarize information, research a subject, or take another approved action.
The agent may connect with email, contacts, financial services, shopping sites, travel tools, messaging apps, and other third-party services.
AI.com says the system can learn from a user’s instructions, choices, and past interactions.
This makes it closer to a digital worker than a basic question-and-answer chatbot.
A chatbot normally explains how to book a trip, while an agent may search for options, open a booking service, enter details, and prepare the purchase.
The user must still give the required permissions, and important actions should be checked before approval.
How do you start using AI.com?
A new user chooses a personal web address and a separate name for their agent.
These addresses use a simple format such as ai.com/name, which makes the agent easy to share or find.
The company says users can create an agent in about 60 seconds without installing special hardware or writing code.
Access starts through a Google account on the sign-up page.
Registration demand may create a queue, so an agent might not become active at once.
AI.com also says usernames are licensed rather than owned, which means the company can reclaim them for inactivity, non-payment, misuse, or brand protection.
Is AI.com free?
AI.com offers a free level and paid Plus and Pro levels, although limits and available features may change while the service remains in beta.
Paid plans renew each month until the user cancels.
Usage depends on both the number of messages and the computing power used by the agent.
According to the user terms, service stops after a user reaches the full limit for a Free, Plus, or Pro plan, then returns at the next billing cycle.
The terms also say that paid fees are generally not refundable and that cancellation takes effect after the current billing period.
Users should check the live price and usage limits before adding a payment method because AI agents can consume more computing power than ordinary chat tools.
Who owns the AI.com name?
Marszalek acquired the AI.com domain in 2025 and built the current service around it.
The Financial Times reported that the domain cost $70 million in cryptocurrency, making it the largest publicly disclosed domain sale at that time.
The short name gives the company a strong advantage because it is easy to type, remember, and trust at first glance.
However, AI.com is not OpenAI, ChatGPT, Google Gemini, or a neutral directory for the whole AI industry.
It is one private company’s product, despite having a domain that sounds like the name of the entire field.
That distinction matters when users decide which company will receive access to their messages, accounts, files, and online actions.
Is AI.com safe to use?
AI.com says conversations, files, and agent data are encrypted and kept private.
Its privacy notice says user environments are isolated through a sandbox model and that their infrastructure and related data are erased after account closure.
The company can still collect account details, device data, usage records, agent interactions, connected-service data, and information needed to run requested tasks.
Some tasks also depend on outside services with their own privacy rules and security risks.
AI.com says it does not use agent interaction or integration data for marketing, but it may use account details and broad feature-use patterns to send promotions.
No online agent is risk-free, especially when it can control apps instead of only producing text.
A careful user should begin with low-risk tasks, connect only necessary accounts, review every permission, and avoid giving the agent control over money or sensitive records until its behavior is well understood.
What makes AI.com important?
The most interesting part of AI.com is not its chat window but its attempt to make computer-using agents simple enough for ordinary people.
Its famous domain and large launch campaign can attract users, but the product will succeed only if its agents complete tasks reliably and explain their actions clearly.
The central test is trust because an agent that can do more work can also make larger mistakes.
AI.com is therefore an ambitious early product with a valuable name, a clear consumer focus, and real promise, but users should treat its beta agents as supervised helpers rather than fully independent workers.