Generated automatically from config-schema.json.
OpenAI API settings.
Your OpenAI secret key.
The language model to use for generating responses. Note that pricing varies between models. See OpenAI's API docs for more details.
Discord bot settings.
Your Discord bot's application ID.
Your Discord bot token.
Your Discord user ID. Only this user is allowed to manage user access, purge the entire database, etc. This user is also always granted access to the bot.
The Discord bot's status. This will be updated every 5 minutes.
The status type/prefix.
Acceptable values: Playing, Listening, Watching
The text of the status.
Placeholders:
{tokens_total}: The total number of tokens used.{tokens_month}: The number of tokens used this month.{price_total}: The total amount of money spent, in USD.{price_month}: The amount of money spent this month, in USD.{interactions_total}: The total number of interactions so far.{interactions_month}: The number of interactions this month.
A list of bot user IDs that are allowed to interact with the bot. This can be useful for automation, but make sure to avoid any recursion!
A set of message objects to send with every request to the language model, before the user's input or context. These messages can be used to influence the behaviour of the model.Each message object should have a role property and a content property.
Messages with a role of system are meant to instruct the model, while messages with roles of assistant or user can be used to start a conversation without the user's input.
The role of the message. Generally, you should only use system messages.
Acceptable values: system, assistant, user
The content of the message.
Placeholders:
{bot_username}: The bot's username{user_username}: The user's username{user_nickname}: The user's server nickname, global nickname, or username{time}: The current system time in 12-hour format (e.g. 7:27 AM){date}: The current date and day of the week (e.g. Wednesday, April 26, 2023){timezone}: The system timezone in a friendly format (e.g. Central Standard Time)
The maximum length in text tokens that a user's input can be.
The maximum length in text tokens that the language model's response should be. Responses exceeding this length will be cut off. Responses exceeding Discord's 2000 character limit will be sent as text files instead.
The price, in USD, of each text token. This is used to display cost values in the stats command.
OpenAI normally advertizes prices per 1,000 tokens, so be sure to divide by 1,000 when setting this value.
Message entries in the database will be deleted when they reach this many days old. Set to 0 to disable automatic deletion.
When true, any user will be able to use the bot. When false, only users allowed through the users allow command or by request will be able to use the bot.
If a message targeting the bot starts with one of these strings, it'll be ignored.
When true, a "Regenerate" button will be included in all language model responses. Responses can also be regenerated from the Apps menu, regardless of this setting.
The number of seconds after which an incomplete request to OpenAI should be terminated. Keep in mind that you may still be charged for requests that time out.
The number times an OpenAI request should be retried if it fails or times out.
HTTP server settings.
Whether or not to enable the HTTP server. Requests to /invite will redirect to the bot's invite URL, and all other requests will redirect to the GiHub repository.
The port to host the HTTP server on, if enabled.
The public hostname that provides access to this HTTP server. This is used to generate invite and conversation viewer URLs.
Whether or not hostname defined above should be referenced using HTTPS, in the event that you're using a proxy.