How It Works
- Users upload documents via the Agentic App Playground or programmatically using APIs.
-
The Platform processes uploaded files in two ways:
- Content Extraction — For supported file types, the Platform extracts content and uses it as context.
-
Metadata Storage — For all supported file types, the Platform captures and stores the following metadata in
sessionMeta.artifacts:Field Description fileIdUnique identifier for the uploaded file typeType of data filenameOriginal filename as uploaded by the user mimetypeMIME type of the uploaded file isActiveWhether the document applies to the current conversation. Learn more. downloadUrlTemporary direct access URL, valid for 30 days from date of upload
- Metadata is accessible in agent and supervisor prompts or code tools, and can be passed to third-party services for further processing. Learn more.
Accessing Uploaded File Metadata
Access all uploaded file artifacts from session memory:downloadUrl of the first uploaded file:
downloadUrl, or pass the URL to a third-party service for further processing.
Example: Insurance Claim Document
A user uploads supporting documents for an insurance claim. The insurer requires the document for verification and future reference.- User uploads the supporting documents for the insurance claim.
- The Platform stores the file and captures its metadata, including the
downloadUrl, in session memory. - A tool shares the file metadata and
downloadUrlwith the insurance system. - The insurance system downloads and processes or stores the file.
- The agent confirms to the user that the document was successfully shared with the insurer.