YAPZEK® Stream
One speech. Every language. Live.
Closed-network, GPU-accelerated multilingual broadcast platform. Speech is transcribed, translated and voiced to every listener in their own language — in real time, without leaving your campus.
YAPZEK® Stream combines a low-latency multilingual translation pipeline with real-time broadcast technology. A presenter speaks; the platform transcribes, translates into the selected target languages and synthesizes natural AI voices — all delivered to listeners through a web page opened with a QR code. No app, no installation.
Everything runs inside the local network. Video and audio pipelines are separated, so subtitles, voice and transcript are fed from the same stream without blocking each other. Recordings, transcripts and archives are stored on the YAPZEK® Storage Server.
- Presenter flow in five steps: source, languages, AI voice, start, share QR
- Single-screen control panel: live status, camera / screen share, transcript, recording, listener management
- Separate language channels, subtitles over video, TXT / JSON export
- Noise suppression, recording and archive on YAPZEK® Storage
- Authorization and access control inside the institution
Closed-circuit / offline
Works on the local network without internet. Data never leaves the campus.
GPU-backed low latency
Real-time speech processing, translation and multi-language broadcast.
Listener picks the language
Scan the QR, choose a language, listen. Live video, subtitles and transcript included.
Scales with you
Single server to multi-GPU, multi-room and multi-campus deployments.
Every patient understood, in their own language.
Hospitals serve patients who speak dozens of languages, and a misunderstanding at the desk, in the examination room or at discharge is a clinical risk. YAPZEK® Stream Healthcare puts real-time, two-way AI interpretation between clinician and patient: the doctor speaks Turkish, the patient hears Arabic, English, Russian or German within a second — and answers in their own language.
The system runs on the hospital's own network. No patient audio, transcript or personal data leaves the campus, which keeps KVKK and GDPR obligations inside the institution. Medical congresses, training sessions and multi-language patient briefings use the same platform in broadcast mode.
- Two-way interpretation: clinician ↔ patient, sub-second, 20+ languages
- Works at the front desk, in exam rooms, emergency and discharge
- Patient joins with a QR code on their own phone — no app, no installation
- Transcript of the consultation can be archived with the patient record
- Fully on-prem: audio and data never leave the hospital network
- Same platform for congresses, training and multi-language briefings
One speech. Every language. Live.
Closed-network, GPU-accelerated multilingual broadcast platform. Speech is transcribed, translated and voiced to every listener in their own language — in real time, without leaving your campus.
- Conference halls
- Universities and lecture theatres
- Public institutions
- Defence and secure networks
- Hospitals and medical congresses
Recording and transcript panel
Lecture hall listeners on their own devices
Presenter view with language channels
- Deployment
- On-prem, closed network, optional offline
- Pipeline
- STT → translation → TTS → subtitle, GPU accelerated
- Listener access
- Web / QR / NFC, no app
- Storage
- YAPZEK® Storage Server, S3 compatible
- Scale
- Single server → multi-GPU → multi-campus
Ready to put AI on your own infrastructure?
Tell us about your environment, your data boundary and your latency target. We will design the pipeline with you and bring it live within days, not months.