I've been using Claude Code, Hermes, OpenCode, Codex and a bunch of other tools pretty heavily. Like a lot of people, I use multiple accounts, a couple GPT subscriptions for heavy coding, Claude for frontend and writing, Gemini for long context, OpenRouter, Cloudflare, NVIDIA endpoints, etc.
The tokens were technically available, but it required constant manual work. Switching accounts, hitting limits mid-session, and babysitting everything got old fast.
So I built ReRouted: a lightweight macOS menu bar app that acts as a local gateway. You point all your tools to one local endpoint and it handles routing and automatic fallback across your accounts.
How it works:
- Connect your accounts (Claude via OAuth, Codex/ChatGPT, Grok, custom OpenAl-compatible endpoints, etc.)
- Create a route (e.g. "coding") with your preferred order
- Use the single local URL + one generated key everywhere
- Access all of your providers and models from a single local endpoint
- If a provider hits a 429, 5xx, timeout, or fails before output starts, it instantly and silently tries the next one in your route
It’s fast, happens in the background, and works incredibly well.
Fully open source and free.
Report
would love to see a per-provider latency and error rate overlay right in the menu bar so I can see which one is actually faster for the task at hand before it swaps automatically
Report
the fact that it just lives quietly in the menu bar and silently swaps providers when you hit a limit is honestly kind of brilliant, the kind of thing you don't appreciate until you've been juggling API keys manually
Report
this is super useful, honestly love the idea of not getting blocked mid workflow. one thing though, would be great if you could set custom thresholds per provider so it switches before hitting the actual limit, basically predicting the cutoff based on usage patterns
Report
The automatic fallback between providers is genuinely clever, especially keeping everything tucked into the menu bar so it stays out of your way. Nice execution on something that could have easily turned into a clunky switcher.
Report
No reviews yetBe the first to leave a review for ReRouted
would love to see a per-provider latency and error rate overlay right in the menu bar so I can see which one is actually faster for the task at hand before it swaps automatically
the fact that it just lives quietly in the menu bar and silently swaps providers when you hit a limit is honestly kind of brilliant, the kind of thing you don't appreciate until you've been juggling API keys manually
this is super useful, honestly love the idea of not getting blocked mid workflow. one thing though, would be great if you could set custom thresholds per provider so it switches before hitting the actual limit, basically predicting the cutoff based on usage patterns
The automatic fallback between providers is genuinely clever, especially keeping everything tucked into the menu bar so it stays out of your way. Nice execution on something that could have easily turned into a clunky switcher.