Frequently Asked Questions

SpokeIt will be released in stages to the world. We plan to improve our beta in English while we work with Smile Train to translate and adapt SpokeIt for other parts of the world. Our first large-scale release is slated for the summer of 2020.

Yes! We are working diligently to improve the SpokeIt beta in various areas. We are running many studies and playtests and would love to work with you! If you would like to get your hands on SpokeIt before release, join our beta list here.

Depending on the answers you provided in the form, we will contact you when we are running a study or playtest that we think you may be interested in.

We will provide details and support when we contact you. Most likely, you will be invited to download the app via an Apple TestFlight link. We hope to work one-on-one with some testers, and in these cases will have SpokeIt on our own hardware.

Our mission is to provide accessible speech therapy around the world to those who need it. There will always be a free version of SpokeIt. The free version will operate using our offline critical speech recognition or a game controller. Unfortunately, our online features that are highly recommended for improved speech recognition accuracy come with operating costs. We are still exploring ways to cover these operating costs without passing them onto our players.

SpokeIt was specifically developed for children born with cleft lip or palate who have had corrective surgery. Soon we will also support children with cleft lip or palate who have not yet received surgery by limiting vocabulary to low-pressure sounds and by guiding correct placement. SpokeIt has had limited testing with Stroke Survivors, adults with Autism Spectrum Disorder, Down Syndrome, Cerebral Palsy, and other developmental disabilities. While our studies show these populations enjoy SpokeIt and find it usable, we need to do more testing on the efficacy of long-term use.

SpokeIt currently “hears” 26 types of speech errors including initial deletions, final deletions, lisps, vowel swaps (fun->fan->fen), and high-pressure omissions (pill->mill, dad->nan). Currently, SpokeIt is trained for articulation errors, but we have begun work on new machine learning models to support resonance errors such as hypernasality, and fluency errors such as finishing complete sentences. It is important to note that SpokeIt should not be used for diagnosis and MUST be used alongside a speech language pathologist. Machine learning models provide guesses and will never be as effective as a trained professional.

SpokeIt is equipped with a telemetry system that records when, how, and for how long SpokeIt is played. All speech targets, speech performance guesses, and sometimes, speech audio are logged. This data is used to provide insights into speech progress and reports for players, Speech Language Pathologists, and family. We also monitor SpokeIt’s use to make usability improvements and better critical speech models.

We want to provide access to speech therapy to anyone who needs it, including those without regular access to the internet. To this end, we have developed two critical speech recognition systems: an offline and online system capable of working in sync. However, we highly recommend using the online features, as they are much more accurate and reliable. We also support a mediated speech therapy experience where game controllers can be used in lieu of our speech recognition to facilitate speech therapy.

We recognize facilitating speech therapy using an algorithm is crucial for those without access to speech therapy but is not the ideal solution. It is often the case that the expertise of the people surrounding the player are more appropriate to facilitate speech therapy—but even then, speech therapy can be tedious and boring. By supporting game controllers, SpokeIt can mediate therapy instead of facilitating therapy. It can provide the speech curriculum, targets, and motivating game content, while a facilitator uses a game controller to grade speech and control game progression.

Yes, in the case SpokeIt’s algorithms are overly critical or not applicable, just connect a game controller to facilitate therapy at an appropriate difficulty. SpokeIt will still provide the curriculum, targets, and motivating game content, but someone else can grade speech and control game progression.

SpokeIt currently supports English. In the coming year, and most likely in this order, we plan to support Hindi, Spanish, Swahili, and Mandarin. We hope our partnership with Smile Train continues to support the inclusion of many languages and cultures around the world.

The primary contact for SpokeIt is the creator Jared Duval, a PhD student at University of California Santa Cruz.

SpokeIt was developed for children born with cleft lip or palate, and therefore, is meant to support children from 3.5 -10, but we encourage anyone who wants to improve their speech to play.

We tried to support anyone who wants to improve their speech, which is why SpokeIt can be played completely hands-free. All content is closed captioned and provides audio visual cues to help players who may have vision or hearing impairments. The handsfree gameplay is designed to support players with motor impairments. We plan to implement haptic feedback, and support for players who are colorblind. All speech targets are illustrated and can be tapped on to hear how it is pronounced to support players who cannot read.

SpokeIt has a growing dictionary of vocabulary. Our artist is busy illustrating new words to support players who cannot read. Our dictionary is automatically filtered to provide targets that are appropriate for each player’s speech goals. SpokeIt supports the sound, word, and sentence levels of speech therapy. Customizations include the number of syllables in words, specific sounds present in words, the placement of specific sounds in words, the level of therapy (sound, word, sentence, pre/post cleft repair), and words to omit.

SpokeIt is being designed to adjust challenge based on performance. However, our algorithm will never be as good as human intuition, which is why we support game controllers. With a game controller, someone other than the player can facilitate speech therapy and progression of SpokeIt to control difficulty that is appropriate for the context.

We are currently implementing support to include multiple players. Whether the other players have a speech impairment or are supporting someone with a speech impairment, we want SpokeIt to work in many contexts.

We have not tested SpokeIt for language learning but are interested in feedback.

We have done limited testing with Stroke Survivors. We have seen some success, but more testing is needed. We have found that some Stroke Survivors struggle with SpokeIt because it can be difficult to repeat words they hear when memory is impacted.

We have not tested SpokeIt with people with Parkinson’s but are interested in feedback.

We have tested SpokeIt with people with Autism Spectrum Disorder, Down Syndrome, Cerebral Palsy, and others, but more testing is needed for long term efficacy. We are interested in your feedback.

Mobile technology is ubiquitous. Many studies have shown that games have the potential to improve learning when compared to traditional learning contexts. Games can provide a motivating context for tedious activities such as repetitive therapy. We encourage you to research the possible benefits games can have on the world. SpokeIt is classified as a serious game for health, which is a game developed to improve the health or wellbeing of players.