bottlecap is a simple Go client for interacting with OpenAI compatible LLM servers run locally or elsewhere. The app is meant to provide a more streamlined experience for interacting with these LLMs in chatbot style via CLI.
Build the binary with make build or dowload it from GitHub Releases
Within the repo, modify the config file accordingly, or create your own config.json file in the same directory as the binary, with the following fields:
{
"url": "https://example.com", # url of the LLM (including port if necessary)
"bearerToken": "<your bearer token>", # bearer token, if required
"model": "<optionally-specify-model>", # name / id of model (such as llama-scout-maas)
"chatInstructions": "You are a helpful assistant.", # system prompt for chat
"imageInstructions": "Describe this image." # system prompt for "image" subcommand
}
Ask away.
NAME:
bottlecap - Provide inputs to the bottelecap application using quotes
USAGE:
bottlecap [global options]
GLOBAL OPTIONS:
--help, -h show help