Language Confidence is a B2B educational technology company focused on providing the best spoken English tools on the market. Our core LCAT technology processes spoken English language and pinpoints precise areas of improvement.
How accurate is your assessment?
Because we deliver assessment on the phoneme (sound) level, our assessment has proven to be more precise and accurate than human graders. Why don't you try it yourself in our demo.
What can we do with the assessment information that your LCAT technology generates?
The most obvious use case for students is real time feedback on their pronunciation. On top of this, we can also use the assessment results for:
• Real-time feedback for students
• Creation of personalised learning scenarios
• Creation of reports for students, parents and teachers
• Proficiency of spoken English testing
How does your API work?
We use an industry-standard RESTful API. Send us your text & audio files (under 15 seconds) and we will send back a detailed phoneme-level assessment in a JSON file, including the top 5 closest per-phoneme match results. Score results are also provided on a per-word and per-sentence basis.
For more technical details, please e-mail firstname.lastname@example.org
Can I request more endpoints to the API?
We currently offer 2 API endpoints: one for IPA and one for Arpabet. If you'd like to talk more about using the production API once you've exhausted your trial plan, please e-mail email@example.com
Do you offer a free trial?
We sure do! See the 'Pronunciation API Trial Plan V2' information here.
How does your assessment work?
We designed our system to be able to emulate a real human tutor's ability to identify granular level pronunciation errors.
Once you send the audio file, with the corresponding transcript via our API, we deliver back a .JSON file with the results.
This .JSON files contains information on the the pronunciation for the student, giving feedback at for each sound, word and for the whole sentence.
Our AI is comparing the incoming student audio to 100% native speaker data. The score that you receive is how close to a native speaker our AI thinks you are.
How secure is Language Confidence?
Our API is secured via 2 methods:
(1) API keys
(2) Short lived oauth2 access tokens.
To ensure your API keys are never misused or stolen we require clients to request access tokens from their backend server using a set of secret client credentials.
All of this information is provided in our developer portal here.
Yes, ironically, we built this as a product a few years ago and decided to put it on the shelf whilst we focused on our core assessment technology. If you’d like to talk to use more about how we can create personalised learning as a service for your students please contact us via firstname.lastname@example.org
What do the scores represent?
We train our system on native speaker data only. The scores that you see as output from our API are how close the AI thinks you are to a native speaker, for each sound. We then average these scores up for the words and sentences. You can change the difficulty of the scoring, making it more or less lenient for your users by using our SDK.
Can you assess anything else, or just pronunciation?
We are currently working hard to provide AI tools for the other metrics of spoken English such as fluency but for now we want to ensure we have the best pronunciation product on the market.