Running Cheshire Cat AI behind a reverse Proxy

Cheshire Cat AI is a production-ready framework designed to facilitate the creation of AI agents capable of advanced conversational interactions. It allows users to train agents using various document formats—such as PDF, TXT, Markdown, JSON, and web pages—and seamlessly integrate them with external APIs and applications. The framework supports both commercial and open-source large language models (LLMs) and embedding models, offering flexibility in model selection. Developers can extend the agents’ functionalities through a plugin system, enabling the addition of custom features and tools.

We are using Cheshire Cat AI extensively to generate AI agents for our clients and ventures, deploying it through its well maintained Docker image. To release it in the wild we can’t use it in its default configuration, because it publishes its service on a standard port without any support for SSL. But SSL is crucial for any web-based service, mainly due to support for Data Encryption and Privacy Protection.

So, we created a specific docker compose file to deploy it behind a Traefik powered reverse proxy. We detailed the process in an article published on Cheshire Cat AI official Blog: you can find it here.