- Twilio transactions save audio recordings of phone conversations
- Use speech to text analyzer api (Google or Siri or popup archive)
- wit.ai or indico parses those audio files to highlight the following:
- Numbers mentioned
- Locations mentioned
- Dates mentioned
- Summary of intent and content
- Possibly people mentioned but it's less important
- Organizes data into a helpful format
- Adds trigger commands that user customizes
- if the trigger words are detected, those commands get added directly to calendar/external location or the words surrounding those trigger words get shown
User: contacts: conversations audio data/file text data locations, people, etc. words surrounding triggers time of conversation triggers
- get/register - registers a user to the app
- post/register
- get/login - logs in and renders the home page that has a list of your stored contacts
- post/login
- post/call - makes a request to twilio, twilio handles the call and records an audio file. We then make a request to an audio to text api. Then run that text through a text analyzer. Add results to the data model.
- give the user a new twilio number
- set the bin number (within TWiML code) to the person they are calling (the contact's number)
- add a recording command within the TWiML code
- or just send a transcription command directly with twilio, and bypass the speech-to-text analyzer
- get/convo - grabs the conversation text sorted by date.
- post/trigger - adding new trigger words or phrases to the data model.