Support
What works and what does not, stated first.
Instead of the response-time commitments and tiered plans a support page usually carries, this one sets out what actually runs today and what is not open yet. We do not publish list prices; deployment scope is decided with you.
The team that built it answers
One team builds and operates the voice engine, console, apps, and backend from one repository. A question reaches the person who wrote that code.
Scoped per deployment
Priced per deployment rather than by a fixed plan or a per-credit meter, starting from your use case, expected volume, and integration surface.
What does not work, we write down
Telephony is built but key-gated, and we hold no certifications yet. Those belong in front of you now, not after a contract.
Controls you can inspect
A fail-closed deploy guard, in-handler authentication, and CI checks on every change — we are happy to walk your security team through them directly.
Questions reach the people who wrote the code
We orchestrate best-of-breed vendor models for speech-to-text, language, and text-to-speech, and build the Korean tuning, the orchestration, and the reliability layer around them ourselves. So a question like "why did it cut the caller off here" is answerable by naming which processor made that call.
The figures run the system; they are not marketing
Time-to-first-audio and response latency are measured on every call. Those numbers exist to run the system day to day and we do not publish them as benchmarks — looking at your own calls' figures during a deployment is more useful than moving a number into a headline.
Reliability
Reliable by design. Not by promise.
A voice call is unforgiving — the failure modes are handled ahead of time, not hoped away. The pipeline is designed not as working-or-erroring but as a system with levels it can step down through.
When one vendor degrades mid-call, the rest of that call moves to another provider and stays there; when one of our own capabilities is uncertain, a flag steps it down. Unverified state fails closed, and a person is the only gate to production.
100+
feature flags that step a capability down instead of dropping the call
Counted in the repository — audit, August 2026.
Implementation state
What is open right now
This is implementation state, not an availability dashboard: whether each capability is live, built but gated, or still in development. How each one is put together is in the documentation.
- Live
- Built and gated
- In development
| Choose a surface | VoiceThe live call path — transports, pipeline, and processors. | ConsoleWhere an operator closes out the work after a call. | AppsThe web and mobile surfaces families and care teams use. |
|---|---|---|---|
| Browser WebRTC calling Starting a call in the browser with nothing to install. | LiveLive | LiveVisible in calls | LiveStarts from web |
| Mobile calling iOS and Android apps reaching the same voice engine. | LiveLive | LiveVisible in calls | LiveiOS · Android |
| Telephony integration The architecture is built but key-gated, and is not running as production phone calls. | Built and gatedKey-gated | Built and gatedPer deployment | Built and gatedPer deployment |
| Live transcription Transcription accumulates while the call runs and is read back afterwards. | LiveUpdates in call | LiveReviewable | Built and gatedSurfacing in progress |
| Recording and call records Recordings and records are stored for later review. | LiveStored | LiveRecord review | In developmentIntegration in progress |
| Custom processors PII redaction, output guardrails, risk scoring, diarization, Korean normalization. | Live30+ processors | LiveResults visible | LiveResults applied |
| Feature-flag rollout Opening capability in stages, and stepping it down when conditions get worse. | Live100+ flags | LiveState visible | LiveStaged exposure |
| Care and manager apps UI is complete; backend integration is in progress. | LiveEngine ready | Built and gated34 domain modules | In developmentUI complete, wiring up |
| Schedules and follow-up Wiring regular calls and post-call work into the operating rhythm. | Built and gatedTriggers ready | LiveHandled in console | In developmentIntegration in progress |
| SOC 2 · HIPAA · GDPR certification We do not currently hold these. What we offer instead is inspectable engineering controls. | In developmentNot held | In developmentNot held | In developmentNot held |
This table does not describe availability or uptime. It states which implementation stage each capability is at; what actually opens for a given deployment is scoped with you during the engagement.
Talk about a deployment
How a deployment gets decided
Instead of a customer story, here is the order scope is actually settled in. A pilot proves one use case against the live stack, and the path to production is drawn from there.
- Starting point
- We start from your use case, expected volume, and integration surface. We don't publish list prices, and pricing is per deployment rather than by a fixed plan or a per-credit meter.
- Pilot
- A pilot proves one use case against the live stack with your team in the loop. This is the stage where what actually works inside a call becomes visible.
- To production
- From there we add concurrent agents, custom processors and guardrails, TTS failover, and feature-flag rollout, and move you onto a production deployment. Telephony, if you need it, is integrated during this process.
We don't claim to own or train the models. What we build is the Korean tuning, the orchestration, and the reliability layer around them.
- custom voice-pipeline processors
30+
custom voice-pipeline processors
- care and consumer backend domain modules
34
care and consumer backend domain modules
Figures are counted in the repository — audit, August 2026.
Working together
From the design through to running it
We shape the call flow and integration path around the environment you already run, and decide together which paths open when — web, mobile, or telephony. It is not a hand-off at the spec; we stay on it until there is a working call.
Console workflows and follow-up handling get tuned to how the work is actually done. If the plan turns out to be wrong, we rewrite it with you rather than billing the difference.
Common questions
Questions, answered
What a Korean-first voice deployment involves, and what it does not. Type a word to narrow the list.
6 questions match.
What do you provide?
An agent that holds a natural spoken conversation in Korean, with an operator console for reviewing those conversations afterward. It runs in the browser and on mobile.
How is a deployment priced?
Per deployment, not by a fixed plan or a per-credit meter. We start from your use case, expected volume, and integration surface, then quote a pilot and a path to production. We don't publish list prices.
Can you connect to our phone system?
Not out of the box. The pipeline ships a telephony-ready architecture, but the transport modules are key-gated and are not running as production phone calls. Telephony is integrated per deployment. Browser and mobile voice are the transports live today.
Which speech and language models do you use?
We orchestrate best-of-breed vendor models for speech-to-text, language, and text-to-speech. We don't claim to own or train them — our work is the Korean tuning, the orchestration, and the reliability layer around them.
Are you SOC 2, HIPAA, or GDPR certified?
Not yet. We do not currently hold SOC 2, HIPAA, or GDPR certifications. What we offer today is inspectable engineering controls — a fail-closed deploy guard, in-handler authentication, and CI checks on every change — and we're happy to walk your security team through them directly.
What happens to conversation data?
PII redaction runs inside the pipeline as one of the processors. Retention, storage location, and access for a given deployment are scoped with you during the engagement rather than assumed.
Not an emergency service
IntuneLabs voice agents are not for emergencies and must not be relied on in one. If you or someone else is in immediate danger, call 119 (fire and medical) or 112 (police) right away.
Still have questions?
Tell us the workflow you have in mind and we will map a path to a working proof-of-concept.
This page carries no response-time commitments, no tiered support pricing, and no uptime figures. Those are values we do not publish — what is here instead is the current implementation state of each capability.
The figures are 30+ voice-pipeline processors, 100+ feature flags, 450+ operator-console components, and 34 backend domain modules — each counted in the repository during the August 2026 audit.