What AI services does Mote use, and who provides them?
Mote uses AI in two distinct ways:
Speech and language services convert between speech, text and languages. They do not create new content.
Generative AI features use large language models, which produce new text in response to a prompt.
This article lists both, along with the provider behind each one.
Speech and language services
These are specialized services. They take what a student says, reads or selects and convert it accurately. Nothing is invented.
| Service | What it does | Providers |
|---|---|---|
| Speech to text | Turns voice recordings and dictation into written text | Deepgram, AssemblyAI, Microsoft Azure, Google Cloud |
| Text to speech | Reads text aloud in a natural voice | Microsoft Azure, Google Cloud, Amazon Web Services, ElevenLabs |
| Translation | Translates selected text into another language | Google Cloud |
Why more than one provider? For speech to text and text to speech, Mote sends each request to a primary provider and automatically falls back to an alternative if that provider is unavailable or does not support a particular requested language. Deepgram is the primary provider for speech to text. Microsoft Azure is the primary provider for text to speech, and supplies Mote's voice catalog.
Generative AI features
These features use large language models. Each one is a specific, bounded task rather than open-ended content generation. Mote does not offer a "write this for me" tool.
| Feature | What it does | Providers |
|---|---|---|
| Dictionary | Produces student-friendly definitions and examples for a word | OpenRouter, routing to OpenAI and Anthropic |
| Vocabulary lists | Generates a vocabulary list for a topic or a piece of text | OpenRouter |
| Sentence Star | Checks whether a student's practice sentence uses a vocabulary word correctly, and gives friendly feedback | OpenRouter, routing to OpenAI and Anthropic |
| Writing Review | Reviews a piece of student writing and suggests improvements | OpenRouter, routing to Google, with Anthropic as fallback |
| Reading text from images | The Mote Chrome Extension has built-in OCR for local text extraction, but this falls back to cloud providers in some cases | Google, with OpenAI as fallback |
What is OpenRouter? OpenRouter is a routing service. Mote sends the request to OpenRouter, which passes it to the language model provider selected for that feature. This lets Mote move between models, or fall back to a second model, without changing the product. The providers reached through OpenRouter today are OpenAI, Anthropic and Google.
A note on reading text from images. This feature uses a general purpose language model, which is why it appears in this section. The task itself is narrow: transcribing text that is already present in the image so that it can be read aloud. The model is not asked to create content.
Which features are available to you
Availability depends on your plan and on the settings your school or district has chosen. Teachers and admins can turn individual features on or off for their students, so not every feature listed here will be active in every classroom.
How these providers are governed
Every provider named above appears on Mote's sub-processor list, together with what it does and links to its terms and privacy policy. Mote vets each provider for use in an educational context before adding it. According to these providers' terms and policies, requests provided by Mote are not used for training purposes.
- Who do you share my data with? is the full sub-processor list.
- Mote's Privacy Policy and approach to student data
- Mote's data security policy
- Where is Mote content stored, and who can access it?
The specific models behind each feature change over time as better options become available. The providers listed here are current. If you need contractual detail about how a named provider handles data, including data processing agreements, contact support@mote.com.