Captive relies on state of the art speech recognition technology, and at the moment this technology runs in the cloud. Most devices are not yet fast enough to recognize text accurately or quickly enough for it to be useful.

Your recorded speech is transmitted directly to Microsoft, and is then transmitted directly back to your device. It is not stored by Captive in a central location, and will therefore only be available on your device.

While Captive tries to use as little bandwidth as possible, if you are running long recognition sessions you should ensure that you are on a network that does not charge for usage.

This is a long topic, but to summarize there are some tips that will help you out:

• Ensure the room is as quiet as possible.
• Have one person speak at a time.
• Speak clearly, but not loudly. Loud speech makes it harder for computers to understand.
• Use a microphone close to the person speaking. When talking to someone with a hearing impairment or deafness, it can be helpful to put wireless earbuds with a microphone (e.g. Apple AirPods) in the ear of the person speaking, since this puts the microphone very close to the most important utterances.

Captive is written and run by a single person, and speech recognition costs money. Until there is a history that shows how much usage to expect from a free trial, or I have a steady income model from Captive, I cannot afford to experiment with free trials, as the costs could very quickly become much larger than I can handle.

I completely understand that trying out a paid solution like this requires trust, which is why I want to disclose this as best I can. The pricing model is designed so that you get plenty of time to experiment with Captive before purchasing.