Putting a browser inside 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 →'s code environment ties their resource use and lifecycles together. Browserbase argues that this becomes an operational problem when browsers need more capacity, crash recovery, or persistent login state independently of the agent. Its architecture proposal separates code execution from the browser runtime, so each can scale on its own and have its own network policy, and browser state can persist without keeping the agent's machine alive. This is a vendor's argument rather than a measured comparison, and the article itself notes that stronger isolation such as a microVM brings extra orchestration complexity.
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.
sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
