No language barrier. Delivered in real time.
We fix what held AI interpretation back: quality and delay. tikaq, real-time AI interpretation captions.
A more immersive conference for everyone.
Date and audience size are enough. We reply within one business day.
Sentence end → caption, average1.5 s
Simon Guerin
APAC Field CTO & Director of Solution Architecture, Mendix
Then we add in a full library of Gen AI building blocks.
01 · sentence end → caption 2.3 s
These are connectors.
02 · sentence end → caption 2.0 s
These are your templates.
03 · sentence end → caption 1.4 s
These are the conversational UI blocks.
04 · sentence end → caption 0.9 s
LLM tools, all plug and play.
05 · sentence end → caption 1.3 s
Then enterprises get turnkey access to large language models and knowledge bases.
06 · sentence end → caption 1.1 s
Directly on Mendix Cloud.
07 · sentence end → caption 1.8 s
Now, this solves one of the biggest blockers that we hear from CIOs.
08 · sentence end → caption 1.5 s
Where do we host our AI safely?
09 · sentence end → caption 2.1 s
Who governs it?
10 · sentence end → caption 1.8 s
Who manages updates?
11 · sentence end → caption 0.8 s
We give you governed enterprise grade homes for that.
12 · sentence end → caption 1.0 s
And now this is the important part because we're not replacing low code here.
13 · sentence end → caption 1.3 s
We're extending on top of it.
14 · sentence end → caption 1.6 s
그리고 여기에 생성형 AI 빌딩 블록 전체 라이브러리를 추가합니다.
이것들은 커넥터입니다.
이것들이 바로 여러분의 템플릿입니다.
이것들이 대화형 UI 블록입니다.
LLM 도구들은 모두 플러그 앤 플레이 방식입니다.
그러면 기업들은 대형 언어 모델과 지식 베이스에 바로 접근할 수 있습니다.
Mendix Cloud에서 바로 가능합니다.
이제 이것이 CIO들에게서 가장 많이 듣는 큰 장애물 중 하나를 해결해줍니다.
우리의 AI를 어디에 안전하게 호스팅할 수 있을까요?
누가 이걸 관리하나요?
업데이트는 누가 관리하나요?
저희가 엔터프라이즈급으로 관리되는 홈을 제공합니다.
그리고 여기서 중요한 점은, 저희가 로우코드를 대체하는 게 아니라는 겁니다.
저희는 그 위에 확장하고 있습니다.
A passage from the Mendix Connect 2025 keynote (November 2025). The caption record of the "Speed first" engine, replayed on its original timing. The engine can be changed to fit the event and the speaker.
5 ways to add real-time translation
Works in any event setup
Example photos are staged. The caption sentences are real output.
Example photos are staged. The caption sentences are real output.
- Sentence end to finished caption (median)
- 1.9 s
- Same, 95th percentile
- 2.9 s
- Continuous run
- 62 min
- Lost sentences
- 0
- Concurrent viewer test
- 6,000 viewers
- Captions delivered
- 100%
- Delivery delay
- about 0.1 s
Caption output, measured
Values from a 12 minute keynote, a 62 minute continuous run and a 6,000 concurrent viewer test, where every viewer received 100% of the captions within about 0.1 s.
Clauses appear first; the finished sentence replaces them once, when the sentence ends.
* Measured under test conditions. Results vary by event and sentence.
See how it worksOne mixer output, captions on every screen
Audience phones, the stage wall, the stream, the operator console and the post-event record, all at once.
01 Now agents aren't just chatbots
02 They are full systems that can take a goal and carry out the steps to achieve it
03 Agents observe, decide and act across your applications
에이전트는 여러분의 애플리케이션 전반에서 관찰하고, 판단하며, 행동합니다.
Agents observe, decide and act across your applications
에이전트는 여러분의 애플리케이션 전반에서 관찰하고, 판단하며, 행동합니다.
Caption sentences are real output from the middle of the Mendix Connect 2025 keynote (November 2025), captioned by the default engine. The sequence and timing of the scene are illustrative.
Built for these events
Overseas keynotes, captioned in Korean
No booth, no receivers. Captions on the audience's phones and the LED wall, with product names learned from the speaker's deck.
- QR on phones, no app
- Chroma key captions on the LED wall
- SRT for the video afterwards
Pricing
Base prices by length. Parallel sessions, output resolution and caption languages change the price, and you can estimate it on the pricing page.
Plan 01
4 hours
KRW 2,000,000
KRW 1,000,000
From one keynote to a half-day seminar.
- Two operators, booth and gear, separate line
- Deck analysis, glossary, rehearsal
- Captions on phones, stage and stream
- CSV, SRT, report
Plan 02
Popular1 day
KRW 4,000,000
KRW 2,500,000
A conference day, a town hall, a product launch.
- Two operators, booth and gear, separate line
- Deck analysis, glossary, rehearsal
- Captions on phones, stage and stream
- CSV, SRT, report
Plan 03
2 days
KRW 5,000,000
KRW 4,000,000
Two-day events, run by the same team and gear.
- Two operators, booth and gear, separate line
- Deck analysis, glossary, rehearsal
- Captions on phones, stage and stream
- CSV, SRT, report
Plan 04
3 days and more
From KRW 5,500,000
Several days, several cities, annual programs. Six days or more is quoted to the schedule.
- Same full package
- Rehearsal for every event
- Quote fitted to the schedule
- One contact throughout
Frequently asked
What if the internet drops?
The operator computer uses its own line, not the venue Wi-Fi, and a standby computer is ready. If the AI connection on the server side drops, it reconnects by itself in about a second.
Which file formats work for the deck?
PowerPoint, Figma Slides, Google Slides, Notion, Word and HWP all work, as does a share link or a PDF export. We only pull product names, people and acronyms into the glossary and do not keep the file.
What about strong accents?
An interview with a South African accent scored the same as a native keynote. The surest check is the rehearsal with the speaker's own voice.
What if a caption is wrong?
The operator adds one line to the glossary and the next sentence uses it. A caption that already went out can be cleared from the audience screens.
How many viewers?
In a 6,000 concurrent viewer test every viewer received 100% of the captions, within about 0.1 s. Audience size does not change the price.
Which languages?
Both directions between Korean and English or other major languages. Viewers pick the language on their phone. Tell us the event languages and we confirm right away.
Is there interpreted audio?
Interpretation is delivered as captions. The speaker's voice goes out through the venue sound as usual, and tikaq shows interpretation captions in sync with it on audience phones, the stage screen and the stream. We do not produce interpreted audio.
Where do the deck and the audio go?
The deck is used only to build the glossary and is not kept. Audio goes to the AI API for transcription and is not recorded. Final captions stay in a database in Seoul and are deleted within five days after delivery.
Leave the captions to tikaq.
Send us the event date and audience size. We reply within one business day. Audience size does not change the price.
Contact us