HernanJiang/CodexRouter

Local-first Codex multi-model router and Windows configurator

What it solves

CodexRouter provides a unified routing layer for the Codex workflow, allowing users to switch between multiple AI model providers, OAuth subscription accounts, and third-party API channels from a single menu. It solves the problem of fragmented access to different AI services by aggregating them into one interface and providing automatic failover between subscription quotas and API backups.

How it works

It operates as a Windows desktop router consisting of a Rust-based console, a compatibility layer (Router Host), and CLIProxyAPI. The tool intercepts requests from Codex and routes them to the configured backend channels. It can merge different providers into a single model ID for seamless continuity or keep them separate for explicit quota selection. It also includes a local SQLite state store and a dedicated dashboard to monitor token usage, account balances, and reset times across all connected services.

Who it’s for

Developers and power users of Codex who use multiple AI subscriptions (like ChatGPT, Claude, Gemini, Grok) and API keys and want to manage them in one place without breaking their conversation context.

Highlights

  • Unified Model Menu: Merge OAuth and API channels into one list within Codex.
  • Automatic Continuity: Seamlessly switch from a subscription model to an API fallback when quotas are exhausted.
  • Aggregated Monitoring: A single dashboard for tracking usage, coding plans, and reset countdowns across various providers.
  • Model-Aware Controls: Independent settings for context windows, compaction thresholds, and reasoning strength per model.
  • Low Overhead: Lightweight tray mode designed for negligible CPU and network activity when idle.

Related

  • Project
  • Project
  • Project
  • Project