patcireamo/ChungusHub
A privacy-first roleplay AI chat app that stores your whole conversation as a branching tree, not a single line
ChungusHub is a frontend for chatting with large language models for roleplay, built to run locally so your data never leaves your machine. Instead of overwriting a chat when you edit or regenerate a reply, it branches, keeping every version, and a story map lets you navigate the whole tree visually. It reads character cards and lorebooks in SillyTavern's existing file formats.
What it does
- Chats are trees: swipes, regenerations, edits and forks all become separate branches instead of erasing what came before, and nothing is lost
- A Story Map shows the whole branching tree on one canvas so you can name branches, mark a canon path and jump to any point
- A built-in assistant can fix character cards, edit chat messages and write lorebooks, with user control over whether it needs approval and what it can access
- A chat memory feature folds old scenes into summaries so long stories still fit within the model's context window, with each branch keeping its own summary
- A prompt builder lets you reorder, toggle and edit every block that makes up the request sent to the model, plus a debug panel showing every request/response with token counts
Why it matters
Because it runs locally and stores data only in a folder on your own machine, it matters to anyone who wants roleplay chat history kept off external servers. Supporting SillyTavern's existing formats also means people can move their characters and lorebooks over without redoing them.
Terms in this repo
- LLM · Large Language Model, an AI model that generates text
- SillyTavern · An existing roleplay AI chat app whose character card and lorebook file formats are widely used
- Lorebook · A file of world info entries that get inserted into the prompt when their keywords appear
- Prompt · The full set of instructions sent to the AI model for a response
- Context window · The maximum amount of text a model can read at once, limiting how much conversation history fits
Repository description (English)
ChungusHub is a local-first, privacy-focused LLM frontend for roleplay.
Open on GitHubTrending repos
- openai/codexOpenAI's coding agent, now runnable straight from your terminal
- cordiverse/cordisA TypeScript framework for snapping feature pieces on and off whenever you need them
- ripienaar/free-for-devA single repo lists every free-tier cloud, API and collaboration tool developers actually need
- Wei-Shaw/sub2apiA relay server that lets one Claude/OpenAI/Gemini/Grok subscription be shared by many users
- multica-ai/andrej-karpathy-skillsA single instruction file that stops AI coding assistants from making silent bad guesses
- eneskirca/nodetermA workspace that puts scattered terminal tabs and AI coding agents on one draggable map
- affaan-m/ECCA toolkit that teaches AI coding agents to work like disciplined engineers
- n8n-io/n8nn8n lets you build automation workflows and AI agents by dragging nodes on a canvas, with code as a fallback
Latest from METAL LAB
- OpenHuman tops GitHub Trending for 9 straight days, a week after launch
- MCP's New Roadmap Enables Stateless, Horizontally Scalable Servers
- Turning Technical PDFs Into AI Skills Cuts Token Use Up to 51x
- Founders are pulling all-nighters to keep up with AI agents running 24/7
- Anonymous model Ox Alpha matches GLM-5.2 on all 60 tokenizer tests