Free LLM Router for OpenAI, Anthropic, and Gemini
Managing multiple API keys for different model providers is a total headache when you're trying to build anything remotely complex. I found this lightweight router that actually lets you switch between OpenAI, Anthropic, and Google (specifically Gemini, not the Vertex mess) through a single interface.
Next
Mojave Paint for macOS Image Editing →
It’s essentially a proxy layer designed to handle the routing logic so you aren't constantly rewriting your integration code every time you want to test if Claude performs better than GPT-4o for a specific task. If you're tired of hardcoding endpoints for every single provider, this might actually save you some sanity.
The setup seems straightforward enough for anyone running a local dev environment or a small production stack. It’s free, which is the only reason I’m even looking at it, but the real value is in the abstraction it provides. You can swap the underlying model without touching your entire backend logic (which is where most bugs hide anyway).
Check out the discussion and implementation details here:
https://news.ycombinator.com/item?id=48900735All Replies （3）
H
humanfeedback
Expert
1h ago
1. Check the latency overhead on the proxy layer. If it's slow, it kills the UX.
0
P
Does it handle fallback logic automatically if one provider goes down? (Don't want another outage disaster).
0
S
I spent a week debugging API rotation issues last month; having one unified endpoint would've saved me.
0