Let's scope
your first use case
Tell us the workflow you have in mind and we will map a direct path to a working proof-of-concept on the live Korean-first stack. One use case, the transports it needs, and what we would measure — agreed before anything is built.
Before you start
Four things worth knowing first
Every engagement is scoped and priced per deployment rather than sold as a plan or metered in credits. Here is what is useful to know going in.
Tuned for Korean from the start
Turn-taking, normalization, and guardrails tuned for Korean end to end — not an English pipeline with a translation layer bolted on.
Transport status, stated plainly
The transports live today are browser and mobile voice. The telephony module is built and key-gated — it is integrated during a deployment rather than switched on by default.
We don't claim the models
We orchestrate vendor models for speech-to-text, language, and text-to-speech. What we build is the Korean tuning, the orchestration, and the reliability layer around them.
Instrumented per call, not published as benchmarks
Time-to-first-audio and time-to-first-byte are measured per call, for running the system day to day.
How it goes
What happens first
Scoping costs nothing — you decide on an engagement after you have seen the plan. What you tell us is used to scope the pilot, nothing else. Read the docs first
Tell us the workflow
A short conversation about the use case, expected volume, and where it runs.
See the shape of it
We come back with a scoped pilot, the transports it needs, and what we would measure.
Run the pilot
One use case, tuned against the live stack with your team in the loop.
Prepare for production
A pilot proves one use case before anything is committed to production. A human is the single gate to production.
What a deployment needs
Every engagement is scoped and priced per deployment rather than sold as a plan or metered in credits. These are the things we work through in the first conversation.
What a pilot needs
What we work through:
- One use case and the conversation it really involves
- An expected call-volume band
- Whether it runs in the browser, on mobile, or both
- What counts as success, and what we would measure
- Your data-handling and retention requirements
Scoping costs nothing
You decide on an engagement after you have seen the plan.
What production adds
What we work through as well:
- Whether telephony is needed — the built, key-gated module is integrated per deployment
- Your governance and deployment requirements — we walk you through the fail-closed model in detail
- Who needs to see what in the operator console
- Which existing systems we need to sit next to
A human approves
A guard wrapper refuses a deploy that would drop a function already running.
What to expect
While a pilot runs, where it has got to stays visible. Below is an illustration of what that progress looks like — it is not real data.
Pilot progress
While one use case is tuned against the live stack, what is finished and what is left stay in view for both teams.
- Workflow mappedThe use case and expected volume are agreed.
- Scope agreedThe transports it needs and what we measure are set.
- Pilot tuningTurn-taking and guardrails are adjusted against real conversations.
- Production reviewHuman approval is required before any production change.
What is built so far
- custom voice-pipeline processors
- 30+custom voice-pipeline processors
- feature flags
- 100+feature flags
- operator-console components
- 450+operator-console components
- care and consumer backend domain modules
- 34care and consumer backend domain modules
Every figure was counted in the repository — audit, August 2026.
Start with your first use case
Tell us the workflow you have in mind and we will scope it with you. We usually reply within a business day.
Contact salesThe panels on this page are illustrations, not real data. Transport availability varies by deployment, and telephony is integrated during an engagement. We do not publish latency benchmarks; per-call measurements are instrumented for operations.