https://vapi.ai/ logo
Join Discord
Powered by
Can we add a path prefix in our S3 bucket?
# support
t
Is there any way to have multiple folders for our different assistants for file upload in an S3 bucket? Can we put a dynamic prefix somehow so that we can distinguish that on our S3 bucket so that we have separate files for separate assistants?
p
Yes that’s definitely possible. You can use dynamic S3 prefixes/folder paths based on the assistant ID, name, or any custom variable so each assistant uploads into its own folder automatically. I can help you set up the bucket structure and upload logic properly so everything stays organized and scalable. @tnigam
s
Hi there, Yes, you can absolutely achieve this. There are two ways to handle folder separation in your S3 bucket:
1. Per-Assistant Folders (Recommended) You can define a specific path for each assistant directly in their artifactPlan. This allows you to distinguish files by assistant name or ID. In the assistant configuration, you can set the recordingPath and loggingPath like this:
Copy code
json
{
  "artifactPlan": {
    "recordingEnabled": true,
    "recordingUseCustomStorageEnabled": true,
    "recordingPath": "https://your-bucket-name.s3.amazonaws.com/assistant-name-a/recordings/",
    "loggingPath": "https://your-bucket-name.s3.amazonaws.com/assistant-name-a/logs/"
  }
}
2. Dynamic Date-Based Prefixes In your global S3 Credential settings, the Bucket Path Prefix field supports LiquidJS templating. This is great for organizing files by date automatically. For example, using {{ "now" | date: "%Y/%m/%d" }} will render as 2026/04/29, creating a nested folder structure based on the day of the call.
Best Practice We recommend combining these: Set the global prefix for date organization (e.g., calls/{{ "now" | date: "%Y-%m" }}/). Use the per-assistant recordingPath to ensure the file lands in a folder specific to that agent. This ensures your bucket stays clean and searchable as you scale!
PreviousNext
Powered by