Browserbase, which provides infrastructure for browser automation, describes an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → design that lets the model write a program to coordinate services. That program can fetch information concurrently, filter large responses and return only the relevant results. This reduces the intermediate material the model must read and the number of separate decisions it must make during a workflow. Browserbase says its internal agent uses this approach. Its engineering account also places permission enforcement outside the generated code: the surrounding system determines which services and actions the program can access.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
