Speech-to-text - Githubissues

ASWE-PDA / Personal-Digital-Agent

3 stars 0 forks source link

Speech-to-text #11

Closed lena143 closed 1 year ago

lena143 commented 1 year ago

As a user, I want the agent to understand what I'm saying so that the agent can work according to my statements.

Acceptance criteria:

[x] API for speech-to-text has been selected
[x] API can be called by an adapter module in the backend

itsmiriam42 commented 1 year ago

APIs for speech-to-text (that also offer t2s):

Microsoft Azure Cognitive Services -> higher accuracy, 5 hours/month free
IBM Watson Speech to Text -> 500 mins (=8.3h)/month free
Google Cloud Speech-to-Text -> most popular, 60 mins free, then $0.006 per 15 seconds

Previous DHBW project documentation using IBM Watson IBM Watson documentation Microsoft documentation

itsmiriam42 commented 1 year ago

used API: Flutter Package speech_to_text

note regarding acceptance criteria "API can be called by an adapter module in the backend": recording of voice will be handled internally, spoken text is available by just using the variable "lastWords" anywhere in the code.