Skip to main content
This document covers advancedmd_get_visit_summary, the native AdvancedMD visit summary tool exposed through the Agents Platform.

​
Authentication

AdvancedMD tools use the existing AdvancedMD connector. Users configure:
  • office_code
  • username
  • password
The client performs AdvancedMD’s two-step API login, discovers the tenant API base URL, caches the security token, and refreshes it when needed. Tools do not accept raw AdvancedMD tokens or direct processrequest.aspx URLs.

​
advancedmd_get_visit_summary

Fetches AdvancedMD visit context for realtime coding by porting Fleming’s GetDateVisitsRaw API call. The tool calls AdvancedMD getdatevisits over the requested date window and returns the target visit plus nearby same-patient visits when available. This tool is read-only.

​
Inputs

Required:
  • visit_id: AdvancedMD visit ID.
Optional:
  • visit_window_days: Days around the current date to inspect for nearby visits. Defaults to 7.

​
Lookup Behavior

The tool:
  • authenticates to AdvancedMD,
  • calls AdvancedMD getdatevisits over the requested window,
  • filters to the target visit_id,
  • returns embedded charge, insurance, and patient data from the visit bundle,
  • returns nearby same-patient visits to support coding context.
This is intentionally not a schedule approximation. It uses the same visit bundle endpoint Fleming uses for coding sync because that payload can include ChargeList, InsuranceList, and PatientList.

​
Output

Common failure reasons:
  • advancedmd_not_connected
  • authentication failure
  • schedule/visit lookup failure

​
Human-readable text

The tool emits markdown for the feed: a ## Visit Summary table, a ## Visit field table, ## Patient, ## Insurance, ## Charges, and ## Nearby Visits sections. The full visit bundle stays in structuredContent.

​
Agent Instruction Guidance

Use this as the first AdvancedMD realtime coding step. AdvancedMD has no chart-alert equivalent and does not run Athena-style precheck; once summary-level skip checks pass, fetch documentation and proceed to the shared coding LLM pipeline.