ballast-ai

Token Station — AI skill for Claude Code

AI community

Local routing control plane for AI agents and LLM providers, with a loopback-only gateway, smart and quota-aware routing, and desktop + CLI apps.

How to install Token Station

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ballast-ai/token-station and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Token Station does

Local routing control plane for AI agents and LLM providers, with a loopback-only gateway, smart and quota-aware routing, and desktop + CLI apps.

Alternatives in AI

Oh My Claudecode — 32 specialized agents and 7 execution modes for Claude Code, with smart model routing and automatic paralleliz 10.9k ★
Claudexor — Multi-harness control plane for Claude Code, Codex, Cursor, and OpenCode: quota-aware rotation across multiple 425 ★
onWatch — Open-source Go CLI that tracks AI API quota usage across 7 providers (Synthetic, Z.ai, Anthropic, Codex, GitHu 397 ★

README

Token Station

**One local gateway for every AI agent.**

Connect Claude Code, Codex, Gemini CLI, Cursor, and other agents to the models you control. Pin a provider, route by task, or use quota before it resets.

[![Release](https://img.shields.io/github/v/release/ballast-ai/token-station?display_name=tag&sort=semver)](https://github.com/ballast-ai/token-station/releases/latest) [![CI](https://github.com/ballast-ai/token-station/actions/workflows/ci.yml/badge.svg)](https://github.com/ballast-ai/token-station/actions/workflows/ci.yml) [![License](https://img.shields.io/github/license/ballast-ai/token-station)](LICENSE)

[Download](https://github.com/ballast-ai/token-station/releases/latest) · [Quick start](#quick-start) · [Docs](docs/README.md) · [Issues](https://github.com/ballast-ai/token-station/issues) · [简体中文](README.zh-CN.md)

Token Station routes each agent to the right model

Highlights

Local by default. The Rust gateway listens on 127.0.0.1:8787 and requires authentication. Agent request traffic leaves the device only when you route it to a cloud provider.
Three routes. Direct pins one provider and model. Smart tiers picks High, Mid, or Low in a single decision. Quota first spends buckets that reset sooner.
Enterprise-managed routing. Enter an enterprise Base URL and credential once. The enterprise service keeps control of its real models and routing policy.
Your providers. Start from 40+ editable presets, add a custom OpenAI-compatible endpoint, or keep work on a local runtime such as Ollama.
Desktop and CLI. Both share the same Rust core. Usage, latency, cost estimates, and request logs st