Skip to main content
This document covers nanonetshealth_get_appointments, the native Nanonets Health appointment-list tool exposed through the Agents Platform.

​
Authentication

The tool uses the Nanonets Health integration. The platform injects:
  • api_key
The API key is excluded from the model-visible tool schema and scopes requests to the connected Nanonets Health organization.

​
nanonetshealth_get_appointments

Fetches lean appointment/encounter records from Nanonets Health. The tool is read-only and supports pagination, date filters, appointment-state filters, facility/provider filters, insurance filters, and free-text search.

​
Inputs

All model-visible inputs are optional:
  • page: One-based page number. Defaults to 1.
  • limit: Page size. Defaults to 50; maximum 100.
  • fetch_all: Fetch and combine every result page. When enabled, page and limit are ignored and the tool uses pages of 100 records.
  • encounter_date_from, encounter_date_to: Encounter date range in YYYY-MM-DD format.
  • coded_date_from, coded_date_to: Coded date range in YYYY-MM-DD format.
  • sync_status, precheck_status, appt_status: Exact status filters.
  • appointment_types: Exact appointment-type filters.
  • facility_codes: Facility/department-code filters.
  • provider_name, insurance_name: Provider and insurance filters.
  • search_term: Free-text search.
  • sort_by, sort_order: Result ordering.

​
Pagination behavior

Without fetch_all, one invocation returns one page and publishes that complete page as one task variable. With fetch_all: true, the tool:
  1. starts at page 1,
  2. requests the maximum page size of 100,
  3. fetches every reported page sequentially,
  4. combines every page’s data array,
  5. returns one structured result and publishes it as one task variable.
The all-pages mode is limited to 100 pages / 10,000 appointments. Results above that limit fail with an instruction to narrow the filters rather than returning silently truncated data.

​
Output

pages_fetched and all_pages_fetched are present only for fetch_all results.

​
Output variables

Every successful result is stored as a JSON-valued task variable. The tool result includes the generated output_variable, and the agent should use that returned reference rather than predeclaring a variable number. Examples:
  • ${VAR_N}.data: all appointments in the result.
  • ${VAR_N}.data[0]: the first appointment.
  • ${VAR_N}.total: total appointments reported by Nanonets Health.
Normal paginated calls produce one variable per invocation/page. A fetch_all call produces one variable containing the combined pages.

​
Agent instruction guidance

Use fetch_all: true when downstream processing must cover every appointment, such as filtering encounters and delegating one child task per valid appointment. Read the actual generated ${VAR_N} from the tool result; instruction text such as “store this in ${VAR_1}” does not create or reserve that variable.