Analysis updated 2026-05-18
Play a simulated character among dozens of AI-driven others that remember past encounters.
Generate an evolving narrative by watching AI characters interact and form opinions.
Experiment with a chosen LLM to drive character dialogue and decisions.
| mistval/yozakura | aredotna/api-examples | ceelog/openweread | |
|---|---|---|---|
| Stars | 15 | 15 | 15 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 2/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Node.js and separate installs for client and server, or the Docker image.
Yozakura is a social simulation game powered by AI. Characters move around a map, talk to each other, and build memories and opinions about one another over time. You can play as one of those characters yourself. The AI drives the dialogue and decision-making for all the other characters, creating a story that changes based on who interacts with whom and what they remember from past encounters. The simulation can support anywhere from a small handful of characters up to dozens or even hundreds, all handled by a language model of your choice. Because the model drives each character's behavior independently, the narrative that emerges is not scripted: it evolves from the accumulation of individual conversations and memories. Installing Yozakura is straightforward. The project offers a downloadable Electron desktop app and a Docker image for users who just want to run it, with links to a getting-started guide in the documentation. Developers who want to dig into the code or contribute can clone the repository directly and run it locally with Node.js. The project includes both a production build mode and a development mode with hot reloading. The codebase is split into a client and a server, both written in TypeScript. Setup requires running an install command in each folder and then starting the server. Once running, the app is accessible in a browser at a local address. VS Code debugging configurations are included in the repository for people who want to step through the code. The project is new and actively developed. Bug reports go to GitHub Issues and feature requests go to Discussions. The author also maintains a Discord server for questions and community conversation.
An AI-driven social simulation game where characters move on a map, talk, and build memories of each other, forming an evolving story.
Mainly TypeScript. The stack also includes TypeScript, Node.js, Electron.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.