Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs-apexspriteai.reliatrack.org/llms.txt

Use this file to discover all available pages before exploring further.

ApexSpriteAI is an AI agent orchestration platform that connects the Claude Code CLI to local large language models running on your own GPU hardware. It gives you a fast, private AI coding assistant with full support for the Model Context Protocol (MCP), so your agent can read files, run commands, and call external APIs — all without sending your code to the cloud.

​
What you can do with ApexSpriteAI

Run models locally

Use LM Studio to serve open-source models like Qwen2.5-Coder-32B on your NVIDIA GPU. Your data never leaves your network.

AI-powered coding

The Claude Code CLI provides an interactive AI coding assistant that understands your codebase and can make changes directly.

Extensible tool use

Add MCP tools to let your agent read files, run shell commands, search the web, and interact with external services.

Remote GPU support

Connect your laptop to a powerful remote GPU server over Tailscale VPN for low-latency inference without local hardware limits.

​
How it works

ApexSpriteAI coordinates four components to deliver AI agent capabilities:
  1. Claude Code CLI — the interface you interact with on your local machine
  2. Tailscale VPN — securely connects your machine to a remote GPU server (optional)
  3. LM Studio — runs the LLM on your GPU and exposes an Anthropic-compatible API
  4. MCP tools — extend the agent’s capabilities with local tool execution
When you send a message to Claude Code, it routes the request to your LM Studio backend, which runs the model and returns a response. If the model decides to use a tool (like reading a file), Claude Code executes the tool locally and feeds the result back to the model.

​
Ready to get started?

Quick Start

Set up ApexSpriteAI from scratch in under 15 minutes.

Architecture Overview

Dive deeper into how the components connect.