Which AI writes the answer. Opus 5 is the most capable (best for competitive strategy, long analysis) and is available to leadership or with your own key. DeepSeek V4 Pro is the standard model for everyone. Sonnet and Haiku are faster and cheaper for quick lookups.

Short — a punchy answer, no headers or tables; good for a quick fact. Medium — the default balance. Long — Jerry uses headers, comparison tables and full frameworks; use it for battlecards, deal strategy and anything you'll paste into a document.

Off by default — Jerry answers from the built-in Streamax knowledge base, which is faster and always on-message. Turn it on when you need something current that the knowledge base won't have (a competitor's new launch, recent pricing, news) and Jerry will search and cite sources. Claude models only; DeepSeek has no web access.

Use your personal Anthropic or DeepSeek key instead of the company key — usage is billed to you, and an Anthropic key unlocks the Claude models. Kept in this browser tab only: never written to a database, a cookie, or the usage log, and cleared when you close the tab.

No key set.

Your sign-in details. Accounts that use an override password (rather than the Streamax mail server) can change it here.