Features
Fix the real bottleneck, and prove you can use the language
Every week you know exactly what to practise and why. Every month you can do something you could not do before — shown by evidence, not by a generic score.
Seven modules that work together
Each module does one clear job. The engine decides what to learn; AI is a tool; evidence is the only thing that counts.
ACC · Onboarding
Onboarding & first plan
Three questions about the work you need to do, a 60–90 second speaking sample, and a first plan in under 5 minutes.
- Sign up with an email code or Google — no password needed
- Goals tied to real situations: meetings, client calls, interviews
- 7 consent purposes you switch on independently, explained before we ask for the mic
- First plan: 1 hypothesised bottleneck + 1 five-minute task
LRN · Sessions
Sessions & fast feedback
10–25 minutes per session. Within 5 seconds of speaking you get exactly one thing to fix.
- 1 error · 1 reason · 1 fix · 1 retry
- At most 2 focus areas per week, with what you do not need to learn spelled out
- Recordings are saved offline and upload when you are back online
- Recovery mode after a break: short sessions, no assessment
EVD · Evidence
Measurement & evidence
Capability is only recognised in new situations, under stated conditions: channel, interlocutor, level of autonomy.
- Layered scoring: voice features → AI → calibrated human raters
- Evidence firewall: practice items are separate from assessment items
- Conditional claims with confidence and an expiry
- Progress page with uncertainty bands, no red scores
TUT · AI tutor
AI tutor
Text chat, push-to-talk role-play, and quota-limited realtime voice for high-stakes missions. Explains in your first language when needed.
- The tutor knows your current task, vocabulary and bottleneck
- Push-to-talk role-play is included in every plan
- Out of realtime minutes? It falls back to push-to-talk instead of blocking you
- AI never writes your progress — only rule-based code decides
PTH · Pathway
Pathway & P50/P80 forecast
A vague goal becomes measurable milestones, then an estimate in weeks based on the minutes you have per week.
- Outcome builder: capability · channel · interlocutor · deadline
- P50/P80 forecast and a feasibility check before you start
- Not feasible? See the cost of each option: more minutes, later deadline, add live
- Weekly tracking: on track / at risk / off track
- Beta
LIV · Live
Live 1:1 & groups
A live session is suggested when it beats AI: a bottleneck AI could not clear, or a mission coming up.
- Teachers get a session brief before the session — 2 minutes to prepare
- Small groups of 3–6 with the same bottleneck, at most 1 band apart
- Post-session observations become real evidence
- Book 12 hours ahead, cancel 24 hours ahead for a credit refund
GAM · Motivation
Sustainable motivation
Rewards real progress, not opening the app. No penalties when you are busy.
- Weekly streaks with an automatic rest day and planned breaks for travel
- Levels tied to real capability: “Ready to call clients”
- Voice time machine: hear your week-one recording next to today's
- Focus mode: switch off XP and badges — keep only progress
Explainable
Every suggestion answers “Why?”
You do not have to trust a black box. Every engine decision carries a reason code, summarised evidence and a review date.
- Why this is this week's focus — with evidence from your own speaking turns
- What you do not need to learn this week, and when it will be reviewed
- You can park a focus or flag a task as too hard or too easy; the engine records it and adjusts
- Every decision is reproducible: same data, same policy version, same result
This week: asking for clarification
Why?
It is what blocks your client calls. In your last 4 turns you went silent for more than 4 seconds 3 times when the client asked a follow-up.
Evidence: 3 of 4 “confirm the order by phone” role-plays, week 2.
Not this week: past perfect — not blocking any task, review in 2 weeks.
Generic courses versus targeting the bottleneck
The same 90 minutes a week, two very different outcomes.
| Criterion | Generic course | Targeting the bottleneck |
|---|---|---|
| What you learn this week | The next lesson in the syllabus, same for everyone | The bottleneck blocking your work, at most 2 focus areas |
| Feedback after speaking | A score or generic praise | 1 main error, the reason and the fix, within 5 seconds |
| Progress is measured by | Hours studied, lessons completed | Verified capability in new situations, with conditions |
| When you are busy | Lost streak, nagging notifications | A planned break, then a short session with no assessment |
| Why you learn this | No explanation | Reason code, summarised evidence, review date |
Get your first plan in 5 minutes
The Free plan is enough to find your real bottleneck. No card required.