LLxprt Code

LLxprt Code is a powerful fork of Google's Gemini CLI, enhanced with multi-provider support and improved theming. We thank Google for their excellent foundation and will continue to track and merge upstream changes as long as practical.

Key Features

Multi-Provider Support: Direct access to OpenAI (o3), Anthropic (Claude), Google Gemini, plus OpenRouter, Fireworks, and local models
Enhanced Theme Support: Beautiful themes applied consistently across the entire tool
Full Gemini CLI Compatibility: All original features work seamlessly, including Google authentication via /auth
Local Model Support: Run models locally with LM Studio, llama.cpp, or any OpenAI-compatible server
Flexible Configuration: Switch providers, models, and API keys on the fly
Advanced Settings & Profiles: Fine-tune model parameters, manage ephemeral settings, and save configurations for reuse. Learn more →

With LLxprt Code you can:

Query and edit large codebases with any LLM provider
Generate new apps from PDFs or sketches, using multimodal capabilities
Use local models for privacy-sensitive work
Switch between providers seamlessly within a session
Leverage all the powerful tools and MCP servers from Gemini CLI
Use tools and MCP servers to connect new capabilities, including media generation with Imagen, Veo or Lyria
Ground your queries with the Google Search tool when using Gemini
Enjoy a beautifully themed interface across all commands

Quickstart

You have two options to install LLxprt Code.

With Node

Prerequisites: Ensure you have Node.js version 20 or higher installed.

Install LLxprt Code:

npm install -g @vybestack/llxprt-code

Or run directly with npx:

npx https://github.com/acoliver/llxprt-code

With Homebrew

Prerequisites: Ensure you have Homebrew installed.
Install the CLI: Execute the following command in your terminal:
```
brew install llxprt-code
```
Then, run the CLI from anywhere:
```
llxprt
```

Common Configuration Steps

Run and configure:
```
llxprt
```
- Pick a beautiful theme
- Choose your provider with /provider (defaults to Gemini)
- Set up authentication as needed

Provider Configuration

Using OpenAI

Direct access to o3, o1, GPT-4.1, and other OpenAI models:

Get your API key from OpenAI

Configure LLxprt Code:

/provider openai
/key sk-your-openai-key-here
/model o3-mini

Using Anthropic

Access Claude Sonnet 4, Claude Opus 4, and other Anthropic models:

Get your API key from Anthropic

Configure:

/provider anthropic
/key sk-ant-your-key-here
/model claude-sonnet-4-20250115

Using Local Models

Run models locally for complete privacy and control. LLxprt Code works with any OpenAI-compatible server.

Example with LM Studio:

Start LM Studio and load a model (e.g., Gemma 3B)

In LLxprt Code:

/provider openai
/baseurl http://127.0.0.1:1234/v1/
/model gemma-3b-it

Example with llama.cpp:

Start llama.cpp server: ./server -m model.gguf -c 2048

In LLxprt Code:

/provider openai
/baseurl http://localhost:8080/v1/
/model local-model

List available models:

/model

This shows all models available from your current provider.

Using OpenRouter

Access 100+ models through OpenRouter:

Get your API key from OpenRouter

Configure LLxprt Code:

/provider openai
/baseurl https://openrouter.ai/api/v1/
/keyfile ~/.openrouter_key
/model qwen/qwen3-coder
/profile save qwen3-coder

Using Fireworks

For fast inference with popular open models:

Get your API key from Fireworks

Configure:

/provider openai
/baseurl https://api.fireworks.ai/inference/v1/
/key fw_your-key-here
/model accounts/fireworks/models/llama-v3p3-70b-instruct

Using xAI (Grok)

Access Grok models through xAI's API:

Get your API key from xAI

Configure using command line:

llxprt --provider openai --baseurl https://api.x.ai/v1/ --model grok-3 --keyfile ~/.mh_key

Or configure interactively:

/provider openai
/baseurl https://api.x.ai/v1/
/model grok-3
/keyfile ~/.mh_key

List available Grok models:
```
/model
```

Using Google Gemini

You can still use Google's services:

With Google Account: Use /auth to sign in
With API Key:
```
export GEMINI_API_KEY="YOUR_API_KEY"
```
Or use /key YOUR_API_KEY after selecting the gemini provider

Managing API Keys

Set key for current session: /key your-api-key
Load key from file: /keyfile ~/.keys/openai.txt
Environment variables: Still supported for all providers

Examples

Start a new project:

cd new-project/
llxprt
> Create a Discord bot that answers questions using a FAQ.md file I will provide

Work with existing code:

git clone https://github.com/acoliver/llxprt-code
cd llxprt-code
llxprt
> Give me a summary of all the changes that went in yesterday

Use a local model for sensitive code:

llxprt
/provider openai
/baseurl http://localhost:1234/v1/
/model codellama-7b
> Review this code for security vulnerabilities

Control and Shortcuts: Settings and Profiles

LLxprt Code provides powerful configuration options through model parameters and profiles:

# Fine-tune model behavior
/set modelparam temperature 0.8
/set modelparam max_tokens 4096

# Configure context handling
/set context-limit 100000
/set compression-threshold 0.7

# Save your configuration
/profile save my-assistant

# Load it later
llxprt --profile-load my-assistant

See the complete settings documentation for all configuration options.

Customizing AI Behavior with Prompts

LLxprt Code features a sophisticated prompt configuration system that allows you to customize the AI's behavior for different providers, models, and use cases. You can:

Create custom system prompts for specific tasks
Override provider-specific behaviors
Add environment-aware instructions
Customize tool usage guidelines

Learn more in the Prompt Configuration Guide.

Next steps

Learn how to contribute to or build from the source.
Explore the available CLI Commands.
If you encounter any issues, review the troubleshooting guide.
For more comprehensive documentation, see the full documentation.
Take a look at some popular tasks for more inspiration.
Check out our Official Roadmap

Provider Commands Reference

/provider - List available providers or switch provider
/model - List available models or switch model
/baseurl - Set custom API endpoint
/key - Set API key for current session
/keyfile - Load API key from file
/auth - Authenticate with Google (for Gemini provider)

Troubleshooting

See the troubleshooting guide if you encounter issues.

Popular tasks

Explore a new codebase

Start by cding into an existing or newly-cloned repository and running llxprt.

> Describe the main pieces of this system's architecture.

> What security mechanisms are in place?

> Provide a step-by-step dev onboarding doc for developers new to the codebase.

> Summarize this codebase and highlight the most interesting patterns or techniques I could learn from.

> Identify potential areas for improvement or refactoring in this codebase, highlighting parts that appear fragile, complex, or hard to maintain.

> Which parts of this codebase might be challenging to scale or debug?

> Generate a README section for the [module name] module explaining what it does and how to use it.

> What kind of error handling and logging strategies does the project use?

> Which tools, libraries, and dependencies are used in this project?

Work with your existing code

> Implement a first draft for GitHub issue #123.

> Help me migrate this codebase to the latest version of Java. Start with a plan.

Automate your workflows

Use MCP servers to integrate your local system tools with your enterprise collaboration suite.

> Make me a slide deck showing the git history from the last 7 days, grouped by feature and team member.

> Make a full-screen web app for a wall display to show our most interacted-with GitHub issues.

Interact with your system

> Convert all the images in this directory to png, and rename them to use dates from the exif data.

> Organize my PDF invoices by month of expenditure.

Uninstall

Head over to the Uninstall guide for uninstallation instructions.

Privacy and Terms

LLxprt Code does not collect telemetry by default. Your privacy is important to us.

When using Google's services through LLxprt Code, you are bound by Google's Terms of Service and Privacy Notice. Other providers have their own terms that apply when using their services.

matthid/llxprt-code