Deepgram global API
Excellent streaming STT models and mature SDKs. Teams still need to consider endpoint geography for Australian callers.
Voxworks hosts a local Deepgram API endpoint for Australian realtime speech-to-text. Build voice agents, live captions, agent assist, and call analytics without sending every stream through offshore infrastructure.
Local route
AU media stream
Caller audio stays on a local network path.
Voxworks STT
Realtime speech recognition patterns.
Your app
Transcript events for actions and analytics.
Speech-to-text providers increasingly compete on streaming behavior, model choice, endpoint options, and deployment geography. For Australian voice products, the endpoint location can matter as much as the model.
Excellent streaming STT models and mature SDKs. Teams still need to consider endpoint geography for Australian callers.
Strong file and realtime transcription options, but migrations often involve different event schemas and realtime session patterns.
Established speech-to-text APIs with streaming modes, custom model choices, and region or deployment tradeoffs.
A local path for developers who want Deepgram-powered realtime transcription with Australian hosting and practical migration help.
The Voxworks SDK is the public integration surface. Under the hood it can preserve familiar Deepgram-compatible streaming semantics while giving developers a Voxworks client, Voxworks auth, and Voxworks event names.
Set region: au, keep your app architecture familiar, and validate supported streaming options during onboarding.
We provision keys for real workloads and confirm rate limits, retention expectations, and security requirements first.
Use the preview to evaluate latency, transcript shape, reconnect behavior, and production monitoring before launch.
import { VoxworksTranscriptionClient } from '@voxworks/transcription';
const voxworks = new VoxworksTranscriptionClient({
apiKey: process.env.VOXWORKS_TRANSCRIPTION_KEY,
region: 'au',
});
const stream = await voxworks.transcription.stream({
model: 'nova-3',
language: 'en-AU',
interimResults: true,
smartFormat: true,
});
stream.on('transcript', (event) => {
console.log(event.text, event.isFinal);
});
stream.sendAudio(audioChunk);Stream live call audio, meeting audio, or media frames and receive transcript events while the conversation is still happening.
Designed for voice agents, call analytics, captions, QA monitors, and agent-assist products where transcript delay affects the product experience.
Keep Australian audio workloads close to your users and product systems, with onboarding support for local network and compliance requirements.
Use familiar model, language, endpointing, interim result, and formatting concepts instead of rewriting your whole transcription stack.
Get help designing reconnect behavior, transcript callbacks, downstream events, and observability for long-running streams.
Discuss private-preview options for ephemeral processing, retention controls, access management, and enterprise security review.
Tell us whether you use the Deepgram SDK, raw WebSockets, a telephony media stream, or another transcription API today.
We confirm models, language settings, formatting options, interim result behavior, timeout handling, and expected concurrency.
Run side-by-side tests on local accents, noisy phone calls, real customer vocabulary, and your downstream transcript parser.
Route pilot traffic to the Australian endpoint, observe latency and transcript quality, then expand as your launch plan matures.
Voxworks is launching an Australian-hosted endpoint for teams that want Deepgram-powered real-time transcription closer to Australian callers and media streams. Contact us for private preview access and endpoint details.
The preview is designed for low-friction migration. Supported clients can point to a Voxworks Australian base URL and use familiar streaming transcription patterns such as /v1/listen, model selection, interim results, and smart formatting where available.
Yes. The first release is focused on real-time streaming transcription for calls, voice agents, live captions, and contact centre monitoring, with interim and final transcript events over WebSocket.
A local endpoint can reduce network distance for Australian audio, simplify data residency discussions, and avoid routing every realtime stream through distant regions before your application can act on the transcript.
Voxworks is preparing a private launch. Teams with Australian realtime transcription workloads can contact us to discuss early access, expected call volumes, security requirements, and migration support.
We work with developer partners with real-time audio workloads, especially voice agents, contact centres, captions, and call intelligence products. Send your use case and we will reply with fit, timeline, and technical next steps.
Expected concurrent streams, audio format, and transcript event needs.
Retention, privacy, access control, and Australian data residency requirements.
Contact Voxworks
Mention the Australian Transcription API and include your current transcription provider if you have one.
Technical context for teams evaluating realtime voice and transcription infrastructure in Australia.