Pausing and resuming conversations in Vapi
# support
Subject: Pausing and resuming conversations in Vapi Hello Vapi team, We are developing an agent that needs to pause for a minute to give the user time to think about an issue before continuing the conversation. We have a few questions regarding the best approach to achieve this functionality: 1. Is there a way to pause an ongoing conversation without ending the call, waiting for a certain duration, and then resuming the same conversation? 2. If pausing is not directly supported, is the recommended approach to end the current call, store the conversation transcript, and then start a new call while passing the previous transcript to maintain context? 2a. If we go with this approach, should we send the entire transcript from the previous call as a single message when starting the new call? 2b. Should this message be sent as a user message, assistant message, or system message? Thank you very much!
Hi @Tianlun, I'm an AI assistant for Vapi. I'll analyze available sources and get back to you shortly.
----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Create Assistant]() * [Get Assistant]() * [List Assistants]() * [Update Assistant]() -----------------------------
### Answer Based on the information available, Vapi does not provide a direct feature for pausing and resuming conversations within the same call. However, there is a concept of "resuming" conversations by starting a new call with the context of the previous conversation. For maintaining context across calls, you could consider storing the conversation transcript from the initial call and then passing relevant portions of this transcript to maintain context for the new call [(1)]() . This could be achieved by incorporating the previous conversation's transcript as part of the new call's setup, possibly in the form of a "system" message that sets the context without being part of the visible conversation flow to the user. Regarding sending the transcript as a single message or as multiple messages, and whether to mark them as user, assistant, or system messages, it would depend on how you want the assistant to interpret this context. If the transcript is meant to provide background context for the assistant's behavior, marking it as a "system" message might be appropriate. This way, it signifies that the message is for internal processing and not directly part of the conversation flow. For detailed implementation guidance, such as how to structure these messages and integrate them into a new call setup, it's recommended to consult the Vapi documentation or reach out to the support for specific guidance. Unfortunately, the provided documents do not cover the exact mechanics of achieving this workflow in detail [(2)]() [(3)]() [(4)]() [(5)]() [(6)](https:
// [(7)]() [(8)]() [(1)]() [(9)]() . Remember, these strategies involve handling and passing potentially sensitive conversation data. Ensure to implement appropriate data handling and privacy measures, particularly if operating in regulated environments or handling personal data.
Got it. Tagging @User for additional help.
Ah thank you, Sahil!