This is a Go program that interacts with OpenAI's GPT-3 API to enable conversations with virtual assistants. The program is capable of running in three different modes:

• single: In this mode, the user communicates with a single virtual assistant (bot) in a back-and-forth conversation. The bot's response is converted to speech via Google Cloud Text-to-Speech API.
• blog: In this mode, the user must specify the bot and a topic. The bot then generates a blog post in Markdown and a YAML metadata file. The generated content is printed on the console.
• team: In this mode, a group of bots maintain a circular conversation. This is useful for simulating a teacher/student, or a team working on a project with different members being different bots. The bot's response is again converted to speech via Google Cloud Text-to-Speech API.

• You need to have a valid GPT-3 OpenAI token.
• You may optionally have access to Google Cloud Text-to-Speech API.

1. Clone this repository to your local machine.
2. Install dependencies by running the following command:

   go get https://github.com/ds0nt/bots
   

1. Set your OpenAI token using the CHATGPT_TOKEN environment variable:

   export CHATGPT_TOKEN=<your_openai_token>
   

2. You can run the program by typing go run main.go <mode> in the terminal.

For example, running the program in single bot mode can be done using the following command:

go run main.go single BotName

3. Navigate to the bots folder to define new bots for use in the program.

In this mode, users can interact with a single bot. The bot's responses will be converted to speech if Google Cloud Text-to-Speech API is available.

Usage example:

go run main.go single BotName

In this mode, the user must specify the bot and a topic. The bot will generate a blog post in Markdown and a YAML metadata file. The content will be printed on the console.

Usage example:

go run main.go blog BotName "Topic"

In this mode, users can interact with a group of bots in a circular conversation. The bot's responses will be converted to speech if Google Cloud Text-to-Speech API is available.

Usage example:

go run main.go team BotName1 BotName2 BotName3

In the above command, BotName1, BotName2, and BotName3 are the names of the bots that will participate in the conversation.

• --no-tts or -T: use this flag to disable text-to-speech conversion. Useful if you do not have access to Google Cloud Text-to-Speech API.

• The getUserInput() function allows users to input multiline text through <<< and >>> delimiters.
• The tts() function converts text to speech using the Google Cloud Text-to-Speech API. If this is not available, the function will do nothing.
• The removeOld() function removes messages from memory if the message count or total message length exceeds a threshold.
• It's a good idea to use direnv to handle your environment variables, including the ChatGPT token.