FAQ
- Contact us for more information or check out our pricing feature here.
- Transcription solution: 60-minutes of audio at no cost. Check out Saigen.ai to start using it right away. We’d love to get your feedback. Submit your feedback here.
- Speech analytics: Our subsidiary, Callbi.io would be happy to answer all your pricing questions.
Yes. For call centre speech analytics, we make our models available through our subsidiary, Callbi.
For more information on speech analytics for call centres, kindly contact info@callbi.io.
Yes. Saigen’s call center models are trained on various South African dialects and accents.
English (South African), English (Australian), English (International), Afrikaans, Sesotho, isiZulu, Setswana and Swahili.
Yes. Saigen’s TTS solution takes text as input and provides speech as output. It can be used for applications such as chatbots, debt collection, etc. This service is currently in the beta phase, but feel free to reach out at info@saigen.co.za.
This depends on the specific application domain. We specialise in developing custom models, and we almost always see significant improvements over general speech recognition systems.
Yes. Our custom large vocabulary models are trained using in-domain data and keywords can be updated at any given point. We provide a solution that is far more bespoke than just a basic “out-of-the-box” model.
English, Sesotho, isiZulu, Afrikaans, Swahili, (Luganda and Twi under development).
All of our transcription models are trained on a primary language as well as the local dialect of English. These models are able to seamlessly transcribe code-switched data between the primary language and the local English dialect. We have a multilingual call centre model available through our subsidiary, Callbi. This model can handle code-switched speech between most of South Africa’s languages.
Saigen’s models are purposely trained on additional noisy data to achieve robustness against background noise.
Yes, we offer real-time transcription solutions. Please reach out to info@saigen.co.za for more information.
Not yet. Saigen specialises in automatic speech recognition (ASR) for use cases including media monitoring, call centre analytics, and general transcriptions. If you have a business-critical translation problem, do reach out to us though!
Saigen’s clients record their own TV and radio stations. Saigen exposes a set of APIs that clients use to upload their recorded media files and download transcriptions. Saigen regularly updates their models by adding custom keywords & pronunciations upon request.
Saigen does not monitor radio or TV stations on behalf of clients but rather exposes a set of APIs that clients use to upload their recorded media files to for transcription downloads.
Yes. Keywords are updated in real-time and are monitored 24/7, 365 days a year.