Introductionv1

Overview

Voxtar is a voice diagnostics API. You submit a few seconds of patient audio; the engine analyzes the acoustic signal for sub-clinical neurological and respiratory voice biomarkers, scores them against the patient's own baseline, and returns structured diagnostic signals you can act on or write back to the EMR over FHIR. Audio is processed and discarded — only derived, pseudonymized features are retained.

apiStandalone API

Submit audio to /analyze from your own systems — batch screening, patient check-ins, or a portal — and write results to the EMR. Start with the Quickstart.

graphic_eqA layer inside your agent

Already run a voice agent? Fork its call audio into /analyze/stream and every conversation becomes a screening. See Add a diagnostic layer.

Request Lifecycle

keyAuth

data_checkValidation

graphic_eqAnalysis

cloud_doneResult

Base URL

All requests go to the versioned base URL below, served over TLS through a global anycast network for the lowest possible latency.

https://api.voxtar.hallelx2.com/v1content_copy

Supported Content Types

Audio Upload

multipart/form-data

For submitting a voice sample to /analyze. WAV/FLAC/MP3/OGG, 16kHz+.

Metadata & Responses

application/json

Standard JSON for patients, results, and every response body. UTF-8 encoded.

Versioning Policy

The Voxtar API is versioned in the path (/v1/). Major versions are non-backward compatible and require code updates; minor versions add functionality without breaking existing integrations.

check_circleBreaking changes are announced 90 days in advance via email.
check_circleDeprecation warnings ship in the Voxtar-Deprecation response header.