TTS-AmazonPolly

Backend application which utilises Amazon's Polly service to convert text into speech

Endpoints

/convert

PUT

Body: Plain text to be converted VoiceId: one of the Amazon Polly Voices

• choose from http://docs.aws.amazon.com/polly/latest/dg/API_SynthesizeSpeech.html Token: Authorization token

Example: curl -X PUT --data "{\"Body\":\"HelloWorld\",\"VoiceId\":\"Geraint\",\"Token\":\"MyToken\"}" localhost:8080/convert

Credentials

Can be found in ftlabs lastpass

building ...locally (a reminder)

• set up golang (with correct folder hierarchy and $GOPATH, etc)
• put the repo code in the correct spot: go/src/github.com/ftlabs/TTS-AmazonPolly
• pull in all the appropriate dependencies: go get -u -v github.com/ftlabs/TTS-AmazonPolly
• build the executable: go install
• ensure the .gitignore file includes .env
• ensure the assorted env vars are in .env
  • AWS_ACCESS_ID
  • AWS_ACCESS_KEY (the AWS secret)
  • TOKEN (for authenticating access to this service)
• actually, since we are not auto-parsing the .env file, manually export all the env vars
• run the executable: $GOPATH/bin/TTS-AmazonPolly

testing

• start your local server
• cd test/command-line
• ./mac.sh (if you are on a mac, or tweak it if on a different OS)
• listen to the TTS

deploying to heroku (a reminder)

• If you have made changes to any dependencies (added, or updated)
  • install and run godep (see https://github.com/tools/godep)
  • commit the changes
• Push to heroku.