Hi Product Hunt, I am the maker of Squirrel. I built Squirrel out of sheer frustration. I was managing multiple AI projects—from various n8n agents to backend business services—and I hit a wall with three massive headaches: Upgrading models was a nightmare. Updating hardcoded model names (like switching from Claude 3.5 to 4) across dozens of repositories took hours, and I always missed a few. I was bleeding money. Provider prices fluctuate, and manually comparing costs to find the cheapest route for every single call was impossible. Debugging was pure guesswork. Without full request and response logs, fixing broken prompts was a shot in the dark. I needed a central control panel, so I built Squirrel—an open-source LLM API Gateway. Instead of hardcoding specific models, you use a virtual name (like my-smart). You change the mapping once in Squirrel, and all your apps update instantly with zero code changes. It also automatically routes requests to the cheapest available provider, handles failover without downtime, and logs every single API call (even streaming responses) so you can see exactly what is happening. It started as an internal tool to save my own sanity and API budget, but I realized the broader developer community could benefit from it. It is fully open-source under the MIT license. I would love for you to try it out. If you have any feedback, feature requests, or questions, please let me know in the comments. I will be here all day to answer them!

Squirrel LLM Gateway

Open-source LLM gateway to auto-route and cut API costs

Open-source LLM gateway to auto-route and cut API costs