For developers who live in the terminal, jcllm offers a convenient and efficient way to access large language models. While web interfaces like ChatGPT are powerful, jcllm provides a streamlined alternative for quick queries.

The other principle of jcllm is simplicity. It is offered as a simple executable for all major operating systems. You do not need Python, NodeJS, or Docker to install it. Just download and run. Try it out and let me know what you think!

Go to https://github.com/jlcheng/jcllm/releases and download the archive for your operating system. The archive will contain a single executable named jcllm.

1. You can run jcllm without any configuration, by specifying required options on the command-line.

jcllm --provider gemini --model=gemini-1.5-flash-8b --gemini-api-key=...

2. You can also set options as environment variables, to avoid specifying the API key on the command-line.

export JCLLM_GEMINI_API_KEY=...

jcllm --provider gemini --model=gemini-1.5-flash-8b

Any CLI argument can be turned into an environmnet variable by 1) Adding "JCLLM_" as a prefix; 2) Making the string uppercase; and 3) Changing dashes into underscores. e.g., --provider becomes JCLLM_PROVIDER and --gemini-api-key becomes JCLLM_GEMINI_API_KEY.

3. You can also use a configuration file.

The program will look for a configuration file in the local directory named .jcllm.toml, or $HOME/.jcllm.d/jcllm.toml.

provider="gemini"
model="gemini-1.5-flash-8b"

gemini-api-key="..."
gemini-models-list=[
  "gemini-1.5",
  "gemini-1.5-flash-8b",
  "gemini-2.0-flash-exp",
  "gemini-2.0-flash-thinking-exp", 
]

openai-base-url=https://api.openai.com/v1
openai-api-key="..."
openai-models-list=[
  "gpt-4o",
  "o1-mini",
  "o1-preview",
]


system-prompt="""
Talk like a pirate.
"""