get_logs

Get event log entries for the patient with optional time, type, trace, and JMESPath projection filters.

Parameters

Parameter	Required	Type	Description
`patient_id`	Yes	`string`	Patient user ID
`date_from`	No	`string` (ISO 8601)	Inclusive lower bound on `EventLog.timestamp`. Date-only values (e.g. `"2018-07-01"`) are localized to start-of-day in `timezone` before converting to UTC.
`date_to`	No	`string` (ISO 8601)	Inclusive upper bound on `EventLog.timestamp`. Date-only values are treated as end-of-day in `timezone`.
`timezone`	No	`string` (IANA)	Timezone for resolving date-only `date_from` / `date_to` (default: `"America/New_York"`). Full ISO timestamps with a time component are honored verbatim.
`limit`	No	`integer`	Maximum number of logs to return (max 200, default: `50`)
`sort`	No	`"desc"` or `"asc"`	Timestamp sort order (default: `"desc"`, newest first; use `"asc"` when scanning from the start of an observation window).
`log_types`	No	`array`	Optional list of `EventLog.type` values to filter by
`trace_type`	No	`string`	Filter by `trace.object_type` (e.g. `"check_in_response"`)
`trace_id`	No	`string`	Filter by `trace.object_id` (opaque id string)
`query`	No	`string` (JMESPath)	JMESPath expression applied server-side to the serialized logs array. The expression root is the list of log entries. List results replace `logs`; non-list results (e.g. `length([])`) appear under `result`. Use to project only the payload fields you need; see examples below.
`format`	No	`"raw"` or `"markdown"`	Output format (default: `raw`). `"pretty"` is accepted as a deprecated alias for `"markdown"`.

`query` examples

# Extract only the Chief Complaint section from clinical notes (drops Medications / HPI / Social History)
[*].{timestamp: timestamp, complaint: payload.sections[?section_label == 'Chief Complaint'] | [0].content}

# Filter to notes and extract Assessment And Plan only
[?type == 'clinical_note'].{timestamp: timestamp, plan: payload.sections[?section_label == 'Assessment And Plan'] | [0].content}

# Project the first lab result per entry
[*].{timestamp: timestamp, test_name: payload.results[0].test_name, value: payload.results[0].value_numeric, unit: payload.results[0].unit}

# Count procedures by name (case-insensitive; see note below)
length([?icontains(payload.procedure_name, 'chemotherapy')])

clinical_note payloads include Medications, History Of Present Illness, and Social History sections that are typically 200–700 characters each and nearly identical across notes. Always pass a query that projects only the sections you need when fetching clinical notes; this can reduce response size by 80–95%.

Case-insensitive matching: the built-in JMESPath contains() is case-sensitive. Because clinical entity names are typically Title-Case (e.g. "Chemotherapy (procedure)"), the server also provides three custom functions in query expressions: lower(str), upper(str), and icontains(haystack, needle): a case-insensitive, null-safe substring test. Prefer icontains(field, 'value') (or contains(lower(field), 'value')) for any name/text filter. These functions are available on both get_logs and get_view_block queries.

Example output

Illustrative: format: raw returns structured records in the logs array; a query replaces logs with the projected result.

{
  "count": 1,
  "logs": [
    {
      "type": "symptom_report",
      "timestamp": "2026-03-18T09:00:00Z",
      "payload": {
        "instrument": "esas_r",
        "symptoms": [{ "name": "fatigue", "score": 6 }]
      }
    }
  ]
}

Getting started

Tools

Advanced

Parameters

`query` examples

Example output

​Parameters

​query examples

​Example output

Parameters

`query` examples

Example output