speech-recognition-bot

README

speech-recognition-bot

Speech recognition bot for Telegram using Yandex SpeechKit API through gRPC.

Currently only the client for the SpeechKit API is implemented (as POC). Bot will be implemented at the later stages.

Installation

This project uses Go modules. To install it, starting with Go 1.12 you can just use go get:

go get github.com/nezorflame/speech-recognition-bot

or

go install github.com/nezorflame/speech-recognition-bot/cmd/speech-client

Also you can just clone this repo and use the build/install targets from Makefile.

Prerequisits

Make sure you have acquired:

• OAuth token
• Folder ID (can be found at your Cloud page after you've selected your project (in the form of https://console.cloud.yandex.ru/folders/YOUR_FOLDER_ID)

Bot

Located at cmd/speech-recognizer-bot. Uses config in the TOML format. Example can be found at config.example.toml.

Usage:

--config string      Config file name (default "config")
--log-level string   Logrus log level (DEBUG, INFO, WARN, etc.) (default "INFO")

Client test app

Located at cmd/speech-client.

Usage:

--audio-file string   Audio file path (for recognition)
--folder-id string    Yandex Cloud folder ID
--lang string         Language to detect (default "en-US")
--log-level string    Logrus log level (DEBUG, WARN, etc.) (default "INFO")
--token string        Yandex Cloud OAuth token