What is awesome-ai-companion?

This repository is a curated list, not a piece of software you install and run. It collects open-source projects related to building a long-term AI companion, meaning a chatbot or assistant designed to remember you, hold a persona, and stay in touch over time rather than reset after every conversation. Entries are organized by what part of a companion system they cover: chat clients, phone-simulator apps, memory systems, voice and visual presence, perception, real-world integrations, and game-like worlds for agents to inhabit. The list is written to help newcomers pick an entry point based on how much setup they want to do. For someone who wants a working companion right away with no coding, it points to phone-simulator apps where you just add an API key. For people comfortable installing an app and editing a config file, it suggests an Android chat client paired with add-on plugins for background messaging and memory. For those who want to run a full companion system on their own server, it recommends a backbone chatbot framework combined with separate memory, voice, and perception components, along with written guides that walk through complete setups. Each entry in the list includes a short description based on the linked project's own README, the primary language, what platform it runs on such as Android, self-hosted, or web, and a status label showing whether it is ready to use, needs adaptation, is a building block for other projects, or has thin documentation that should be double checked before relying on it. This is best suited for people already experimenting with AI companions or chatbots who want to survey the existing open-source landscape before choosing tools, rather than for someone looking for a single ready-made product. Since it is a list of links rather than a running application, there is nothing to install from this repository itself. Anyone using it will need to click through to each linked project and follow that project's own setup instructions. The full README is longer than what was shown.