LLaMB is a command-line tool that allows you to interact with Large Language Models directly from your terminal. It features conversation history, file input/output, and support for multiple LLM providers.
LLaMB is a command-line tool that allows you to interact with Large Language Models directly from your terminal. It features conversation history, file input/output, and support for multiple LLM providers.
npm install -g llamb
Or clone this repository and run:
npm install
npm run build
npm link
You can set your OpenAI API key in your environment:
export OPENAI_API_KEY=your_api_key_here
Or add a provider using the interactive command:
llamb provider:add
LLaMB supports many providers out of the box with pre-configured settings:
LLaMB securely stores your API keys in your system's secure credential store:
This means your API keys are never stored in plain text in configuration files.
You can update API keys for existing providers:
llamb provider:apikey
If you install from source, you can install the man page with:
sudo npm run install-man
LLaMB comes with a comprehensive man page that you can access after installation:
man llamb
llamb what is the best command to install docker on ubuntu
llamb -f script.js "Explain this code and suggest improvements"
llamb "Summarize this document" -f document.txt
You can save responses to files in several ways:
# Prompted for filename (default: llamb-response-<timestamp>.txt)
llamb "Generate a React component" -o
# Save to specific file
llamb "Generate a JSON configuration" -o config.json
# Overwrite existing files without prompting
llamb "Generate a CSS file" -o styles.css --overwrite
If a file already exists, you'll be asked whether to:
file-1.txt)LLaMB provides an interactive chat mode that allows you to have back-and-forth conversations with the model:
llamb -c "Tell me about TypeScript"
After receiving the initial response, you'll be prompted for follow-up questions:
🦙 Follow-up question (type /exit to end conversation):
You can type your follow-up questions, or use special commands:
/exit or /quit - End the conversation and exit/clear - Clear conversation history/new - Start a new conversation/history - View conversation history/file - Attach a file to your next question/unfile - Remove the attached file/help - Show available commandsLLaMB keeps track of your conversation history to provide context for follow-up questions.
llamb /history
# or
llamb context:history
llamb /clear
# or
llamb context:clear
llamb /new
# or
llamb context:new
llamb -n "What is 2+2?"
# or
llamb --no-history "What is 2+2?"
LLaMB maintains separate conversation histories for each terminal window. This means you can have different conversations going on in different terminal sessions.
To see information about your current terminal session:
llamb /debug
# or
llamb context:debug
llamb models
# or
llamb /models
llamb providers
# or
llamb /providers
llamb provider:add
# Interactive mode
llamb provider:edit
# Non-interactive mode
llamb provider:edit --name openai --url https://api.openai.com/v1
llamb provider:edit --name openai --model gpt-4o
llamb provider:edit --name openai --key YOUR_API_KEY
# Interactive mode - select provider and confirm
llamb provider:delete
# Non-interactive mode
llamb provider:delete --name openai
# Skip confirmation prompt
llamb provider:delete --name openai --force
llamb provider:default
llamb model:default
# or
llamb /model
You can also change the model for a specific provider:
llamb model:default -p openai
llamb -m gpt-4 -p openai how do I install nginx on Ubuntu?
llamb -u http://localhost:8080/v1 what is the meaning of life?
LLaMB provides different UI rendering options to handle different terminal environments:
LLaMB uses a React-based terminal UI powered by ink, which provides a clean, artifact-free interface with proper rendering:
# ink is the default, so no flag is needed
llamb "What is the history of the internet?"
# You can explicitly enable it with
llamb --ink "What is the history of the internet?"
In some terminal environments, you might want to fall back to traditional rendering:
llamb --no-ink "What is the history of the internet?"
If you're experiencing issues with scrollback artifacts in your terminal:
llamb --progress-only "What is the history of the internet?"
# Set ink UI as default (already the default)
llamb config:progress-mode --ink
# Use traditional rendering with real-time streaming
llamb config:progress-mode --disable
# Use progress-only mode (no streaming)
llamb config:progress-mode --enable
# View current settings
llamb config:progress-mode
llamb -m llama2 -p ollama -f code.py -o analysis.md "Analyze this code"
LLaMB allows you to save and reuse common prompts to streamline your workflow.
llamb prompt:list
llamb prompt:add
This will prompt you for:
# Use with a file
llamb -t summarize -f document.txt
# Use with direct input
llamb -t code-review "function add(a, b) { return a + b; }"
# Combine with other options
llamb -t translate -p openai -m gpt-4 -o spanish.txt "Hello world"
llamb prompt:edit
llamb prompt:delete
llamb prompt:show summarize
Prompt templates can include special placeholders that get replaced when used:
{input} - Gets replaced with the command-line question text{file} - Gets replaced with file contents when using -f flag{filename} - Gets replaced with the filename when using -f flagExample prompt template with placeholders:
Please analyze the following code from {filename}:
{file}
Focus on these aspects:
1. Performance
2. Security
3. Readability
Additional notes: {input}
When used as llamb -t analyze -f script.js "Focus on error handling", the placeholders will be replaced with the appropriate values.
When saving responses to files, LLaMB automatically detects the content type and applies appropriate file extensions:
# When asking for code and saving with -o flag (without extension)
llamb "Write a Python function to calculate fibonacci" -o fibonacci
# Will save as fibonacci.py if response is only code
# When providing a filename without extension
llamb "Write a JavaScript React component" -o component
# Will save as component.jsx if response is purely JSX code
This feature:
MIT
FAQs
LLaMB is a command-line tool that allows you to interact with Large Language Models directly from your terminal. It features conversation history, file input/output, and support for multiple LLM providers.
We found that llamb demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
Multiple high-impact npm maintainers confirm they have been targeted in the same social engineering campaign that compromised Axios.
Security News
Axios compromise traced to social engineering, showing how attacks on maintainers can bypass controls and expose the broader software supply chain.
Security News
Node.js has paused its bug bounty program after funding ended, removing payouts for vulnerability reports but keeping its security process unchanged.