Browserbase's Stagehand can avoid asking a LLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition → to solve the same browser action repeatedly. After a successful action, it saves the selector, the information used to locate the relevant page element, along with supporting metadata. On a later request, it checks the action inputs and a fingerprint of the page before reusing that selector. A mismatch sends the action through the normal model-assisted path. The caching explanation shows why repeated forms and stable page templates benefit most: the browser still performs the action, but a matching request can skip another model call.
Terms in this piece · Glossary
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
