Sonde Health API Platform Documentation
Getting Started - Sonde Passive SDK, Cued SDK, and Cloud API
The following is guide to integrate Sonde’s SDKs and APIs in your mobile and web applications.
Should we use SDK or Cloud API?
The following are some tradeoffs for you to select between multiple options. We strongly recommend using Sonde’s Passive SDK as this has many benefits to your end users including security and privacy, lower cost (as no backend resources are needed), offline use, and higher performance.
| Sonde Passive SDK | Sonde Cued SDK | Sonde Cloud API |
---|---|---|---|
Definition | Your user voice will be passively analyzed in the background. Your application will continuously analyze the audio in background and only filters the authorized user voice and rejects the other user voices. The audio will be analyzed on the device and immediately deleted. This offers high security and privacy as well as low latency and does not require internet access. | Your user will be cued (asked a question or an activity to elicit a voice response). Your application will capture the audio on the device. The audio will remain on the device and score will be calculated on the device. This offers high security and privacy as well as low latency and no internet access. | Your application will capture the audio on the device and send it via API calls to the Cloud. Here the score will be calculated and through an API call you will receive the score. |
OS support | Support for both Android and iOS. | Support for both Android and iOS. | Mobile phones and devices with Android and iOS that have audio capture capability. |
Health check support (refer to documentation for more details) | Mental Fitness | Respiratory | Respiratory |
Network connectivity | Application can execute with no network connectivity. However, for the purposes of registration and metering some connectivity is recommended. | Application can execute with no network connectivity. However, for the purposes of registration and metering some connectivity is recommended. | Medium to strong connectivity preferred. Application will still work in slow connectivity but scores return may take up to 30 seconds. |
Latency | 4 to 6 seconds based on the mobile device. | 2-4 seconds based on mobile device speed and length of audio captured. | 12-30 seconds based on the network connectivity and length of audio captured. |
Data Transfer | Optionally scores, and device specific anonymized meta-data may be uploaded to the Cloud. | For billing and building personalized models, usage metering and meta-data is uploaded to the backend. Data includes scores, date/time stamp, users YOB, gender scores, and device OS and model. | Real-time audio uploaded to generate the scores. 600 KB for Respiratory audio uploaded (6 seconds of ‘Ahh’); |
Security and Privacy | Analyzed audio for scoring is not stored on the device or in the Cloud. | Captured audio for scoring is not stored on the device. App developers need to delete the audio after getting scores. | All audio/data transfer to the cloud is on HTTPS. |
Integration effort (not including UI/UX) | 3-5 days | 1-3 days | 1-3 days |
Upgrades and bug fixes | SDK upgrades will require app upgrades. | SDK upgrades will require app upgrades. | Higher flexibility as this does not require changes in your Edge app build. |
Access to API/SDK | Contact Sonde | Contact Sonde | Contact Sonde |
Cost | Contact Sonde | Contact Sonde | Contact Sonde |
For additional information, please contact support@sondehealth.com.
Sonde Health