Tinkoff Speech API Examples

Usage

Clone this repo

$ git clone --recursive https://github.com/TinkoffCreditSystems/tinkoff-speech-api-examples.git
$ cd tinkoff-speech-api-examples

Install requirements

$ pip3 install -r requirements.txt

Generate protobuf and grpc definitions for your language (Python):

$ ./sh/generate_protobuf.sh

Setup environment

Set STT_TEST_API_KEY and STT_TEST_SECRET_KEY environment variables to your API key and secret key to authenticate on server:

export STT_TEST_SECRET_KEY="SECRET_KEY"
export STT_TEST_API_KEY="API_KEY"

Run examples

$ ./sh/recognize.sh

$ ./sh/recognize_rest.sh

$ ./sh/recognize_stream_file.sh audio/sample_1.mp3

One should install sox library to recognize recording from a microphone (apt install sox / brew install sox):

$ ./sh/recognize_stream_mic.sh

$ ./sh/synthesize_stream.sh

You may get scope tinkoff.cloud.tts is not supported error if your API key does not
support speech synthesis.

Note on endpoint format

You may use stt.tinkoff.ru:443 and tts.tinkoff.ru:443 interchangeably for both speech recognition and speech synthesis.