Openi on Thursday issued a new artificial intelligence (AI) agent, which will be covered in the company’s chatboat. Dodged GPT agent, it is a common purpose agent who browses the web and finds information as well as receiving its virtual computer for an integrated development environment (IDE) for coding. The AI firm based in San Francisco said that the Chat GPT agent is primarily an unity and deep research event of the operator agent. The new agent’s capacity is available to users paid within the Chatboat Web and Desktop Interface.
Chat GPT Agent can complete work on your desktop
The AI firm directly unveiled a new AI agent in the series. It is the company’s third agent offering (after deep research and operator) and the second independent agent. Open said that the Chat GPT agent comes with the ability to interact with websites, deep research skills in information synthesis, and the chat GPT’s intelligence and communication flow. “
Chat GPT can now work for you using your own computer.
Introduction to Chat GPT Agent-Operators action remote browser, Deep Research web synthesis, and a united agent system that combines Chat GPT conversation powers. pic.twitter.com/7un2nc6nbq
– Openi (@Openi) July 17, 2025
The Chat GPT agent can now engage with websites and click on buttons, filter the content, and just do actions such as collecting the required information. Since the agent is integrated inside the chat boot, users can go to request specific steps without leaving chat. It also allows the agent to get more context.
The company said in a blog post that the agent comes with a suit of tools that it uses to perform autonomous acts. It has a visual browser that the agent uses to interact with the web, to submit access to a text -based browser immediately, a terminal, and direct application programming interface (API).
Open said that Chat GPT agent can also use the connector to connect third -party apps such as Gmail and Gut Hub. If an agent needs to be logged in to a website, the user can handle its certificates from the agent and add its certificates. In particular, all these tools are provided to the agent via virtual computer. When it works at work, users can see a small window in the chat interface to understand what it is.
The purpose of this capacity is to handle more complex tasks than previously issued agents. The AI firm highlights that users can add new data to Chat GPT agents, to convert screenshots or dashboards into presentations with editable vector elements, reset meetings, program planning and booking locations, to add new data to the existing spreadsheet without losing formatting. The agent can create a preliminary retirement plan through local tax laws and personal investment strategies, planning and book travel travel, parties, and even experts, and appointments of books.
Chat GPT Agent is currently available to Plus, Team, and Pro users. Although supporters have already had access to this, others are expected to get it in the next few days. In addition, in the coming weeks, the agent will be extended to the enterprise and education level. In particular, Pro users have a monthly rate of 400 messages, while the second tier has 40 messages monthly.
This feature will not be available in the European Economic Area (EEA) and Switzerland, but Open said it was working to enable it in the region as well. With the release of the Chat GPT agent, the AI firm is also planning to close the operator in the coming weeks.
