TTS-service

HTTP microservice using Axum to generate TTS from an HTTP reqwest.

Modes

• eSpeak - Local TTS, low quality. Returns WAV audio.
• gTTS - Cloud TTS, medium quality. Returns MP3 audio
• gcloud - Google Cloud TTS, high quality. Returns OPUS audio. Requires a gCloud API key
• Polly - Amazon Polly TTS, high quality. Returns OggVorbis audio. Requires Amazon Polly credentials

Supported endpoints:

• GET /tts?text={CONTENT}&lang={VOICE}&mode={MODE}&speaking_rate={SPEAKING_RATE}&max_length={MAX_LENGTH}&preferred_format={PREFERRED_AUDIO_FORMAT} - Returns the audio generated.
• GET /voices?mode={MODE}&raw={BOOL} - Returns the supported voices for the given mode as either a JSON array of strings, or a raw format from the source with the raw set to true.
• GET /modes - Returns the currently supported modes for TTS as a JSON array of strings.

Error Codes:

Non-200 responses will return a JSON object with the following keys:

code - int

• 0 - Unknown error
• 1 - Unknown voice
• 2 - Max length exceeded
• 3 - Speaking rate exceeded limits, see the display for more information
• 4 - AUTH_KEY has been set and the Authorization header doesn't match the key.

display - str

A human readable message describing the error

Environment Variables (default)

• BIND_ADDR(0.0.0.0:3000) - The address to bind the web server to

• LOG_LEVEL(INFO) - The lowest log level to output to stdout

• AUTH_KEY - If set, this key must be sent in the Authorization header of each request

• REDIS_URI - The URI of a redis instance to cache requests with

• CACHE_KEY - Fernet encryption key to use to encrypt audio data

gTTS Required

• IPV6_BLOCK - A block of IPv6 addresses, randomly selected for each gTTS request

gCloud Required

• GOOGLE_APPLICATION_CREDENTIALS - The file path to the gCloud JSON

Polly Required

• AWS_REGION - The AWS region to use

• AWS_ACCESS_KEY_ID - The AWS access key ID

• AWS_SECRET_ACCESS_KEY - The AWS secret access key

Docker build variables (default)

• MODES(espeak) - A comma separated list of modes to support