To answer your questions: - The 2000‑token figure shown in the UI is the default maximum combined size of the prompt + model response for the selected model on the current plan. It is not a hard technical ceiling of the model itself, but a plan‑level quota. - You can increase it by upgrading plans or purchasing the addon for AI-Spend! - Hard technical max – each model has its own intrinsic context window (e.g., 8192 tokens for GPT‑3.5, 32768 tokens for Gemini‑1.5‑pro). Even with unlimited quota, you cannot exceed the model’s native limit. - There is no native support for pausing a specific conversation. - For transferring a conversation to a human you can follow the guides here: > 1. https://www.botpress.com/docs/integrations/integration-guides/hitl#human-in-the-loop-hitl > 2. https://www.botpress.com/docs/get-started/manage-your-agent/human-handoff