OpenAI Operator

OpenAI Operator is OpenAI's browser automation agent for people who want ChatGPT to navigate websites, fill forms, compare information, and prepare outputs from natural-language instructions, with access in this record starting at ChatGPT Plus for $20/month and higher limits or shared use listed for Pro and Team plans. It is best understood as a supervised browser agent rather than a fully unattended production automation system: the user describes the goal, ChatGPT opens and interprets web pages visually, and the agent performs clicks, typing, scrolling, tab switching, and information gathering while leaving sensitive or consequential steps under user control.

OpenAI introduced Operator in January 2025 as a research preview for ChatGPT Pro users in the United States. In mid-2025, OpenAI introduced ChatGPT agent mode, which brought Operator-style browser control together with ChatGPT's research, analysis, code execution, and file-generation capabilities. The practical result is that users describe a web task in natural language, then supervise ChatGPT as it opens pages, reasons from screenshots, enters information, handles ordinary website navigation, and pauses for passwords, payment details, final confirmations, or other sensitive actions.

Five concrete facts define the current listing: Operator was announced in January 2025; the original preview was tied to the $200/month ChatGPT Pro plan; ChatGPT Plus is listed at $20/month; ChatGPT Team is listed at $25/user/month billed annually or $30/user/month billed monthly; and the official sources for this record include OpenAI's Operator launch post, ChatGPT agent launch post, ChatGPT agent help article, and Computer-Using Agent post. The pricing data also lists ChatGPT Pro at $200/month and Enterprise as custom pricing, so the economic fit depends heavily on whether the buyer already values ChatGPT's broader assistant, research, coding, and file-generation capabilities.

Operator is useful for occasional workflows on websites without APIs, such as gathering competitor data, booking appointments, building shopping carts, and completing long forms. A researcher might ask it to visit 10-20 competitor pricing pages, collect plan names, note feature differences, and summarize the findings. A small business user might have it check portals, download documents, or enter information into sites that do not offer clean integrations. A consumer might use it to assemble an online grocery order or search appointment availability, then take over before payment or final submission.

Its limitations are material. Screenshot-based interaction is slower than DOM-based scripting, and the agent can misread controls, misclick, follow an unexpected path, or get blocked by authentication prompts, CAPTCHAs, anti-automation systems, complex checkout flows, or high-impact sites that require close supervision. No official source in this record confirms native MCP support or direct API access to the same Operator product experience for custom developer applications. For repeatable production workflows, Playwright, Selenium, Browserbase, UiPath, or a custom automation stack may be more predictable. Operator is strongest when ease of setup, natural-language control, and human supervision matter more than speed, determinism, or large-scale unattended execution.