Contact A-LUX

We reply within 15 minutes

Вы человек? 3 − 1 =

Or contact directly

Speech-to-Text & TTS Engineering

Speech Recognition & Synthesis Voice That Your Business Can Read, Search and Speak

Prices and terms valid as of June 2026

A-LUX designs and ships production speech systems for Kazakhstani and international companies: call transcription, voice commands, real-time speech synthesis, and deep integration with your CRM, call center, 1C and WhatsApp Business. Russian, Kazakh and English out of the box.

WhatsApp
  • Call transcription with speaker diarization and keyword analytics
  • Voice commands and IVR that understand Russian, Kazakh and English
  • Natural TTS voice synthesis wired into CRM, 1C and WhatsApp Business

A-LUX is a Kazakhstan-based digital agency that builds speech recognition and synthesis systems — call transcription, voice commands, and text-to-speech — with full integration into CRM, call centers, 1C and WhatsApp Business. Projects start from ₸1,000,000. Call +7 705 966-25-25.

FAQ

Answers to common questions before ordering from A-LUX.

Does your speech recognition support the Kazakh language?

Yes. We build and tune speech-to-text for Russian, Kazakh and English, including mixed Russian-Kazakh speech that is common in everyday business calls in Kazakhstan. We fine-tune models on your own recordings to handle local accents, names and industry vocabulary.

Can you transcribe our call-center recordings automatically?

Absolutely. We process 100% of your calls with speaker diarization, timestamps, keyword tagging and sentiment scoring, then push the transcript and a short summary straight into the contact card in amoCRM, Bitrix24 or your custom CRM.

How much does a speech recognition project cost?

Projects start from ₸1,000,000. The final figure depends on languages, daily audio volume, required accuracy, and the depth of CRM, 1C and call-center integration. We give a fixed estimate after a short discovery call.

Can the system run on-premise instead of the cloud?

Yes. For banks, fintech and companies with strict data-residency rules we deploy the recognition and synthesis engine fully on-premise or in a private cloud inside Kazakhstan, so audio and transcripts never leave your infrastructure.

Do you integrate voice with 1C and WhatsApp Business?

We do. Dictated orders and tasks flow into 1C as structured documents, and incoming WhatsApp Business voice messages are transcribed and can be answered with synthesized voice notes — turning voice into a fully automated channel.

How natural does the synthesized voice sound?

We use modern neural TTS that sounds close to a real person, with correct intonation in Russian, Kazakh and English. If you need a recognizable brand voice for IVR or notifications, we can create a custom voice as part of the project.

How long does it take to launch?

A typical production system takes 6–12 weeks: discovery and accuracy goals, model selection and tuning on your data, integration with your CRM/PBX/1C, a pilot on live traffic, and hardening for full volume.

How accurate is the recognition?

On clean telephony audio we typically reach 90–95% accuracy after tuning on your domain vocabulary. We define a target word-error-rate up front, measure it on your real recordings, and keep improving the model with new data after launch.

WhatsApp

Get a speech technology estimate

Fill in the form — our manager will reply within 15 minutes. Free consultation and quote.

By submitting you agree to the processing of personal data