donbananas/multiagent-brainengine-sillytavern — explained in plain English
Analysis updated 2026-05-18
Give AI roleplay characters a simulated inner life with emotions, memory, and physical reactions.
Have characters follow a daily schedule so conversations feel grounded in a life outside the chat.
Build a custom narrator character that sees every character's hidden thoughts to manage scenes.
| donbananas/multiagent-brainengine-sillytavern | 100/awesome-machine-learning | adam-s/car-diagnosis | |
|---|---|---|---|
| Stars | 8 | 8 | 8 |
| Language | Python | Python | Python |
| Last pushed | — | 2024-08-07 | — |
| Maintenance | — | Stale | — |
| Setup difficulty | moderate | easy | moderate |
| Complexity | 3/5 | 1/5 | 3/5 |
| Audience | general | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires an existing SillyTavern install, your own AI provider API key, and specific SillyTavern settings changes to work correctly.
This project is a Python proxy server that plugs into SillyTavern, a tool people use for AI powered character roleplay, and gives its characters a much richer inner life. Instead of one AI call replying to a message, it splits a character's response into six separate AI agents that each handle a different part of thinking. One agent works out immediate physical reactions like a racing heart or tensed muscles, another tracks the character's underlying goals and emotional bonds, another filters the conversation through the character's past memories and worldview, and another tries to read between the lines of what you actually mean, not just your literal words. There is also a small amount of built in randomness meant to make thoughts feel less mechanical. A separate agent gives each character an actual daily schedule and weekly calendar, so if you are chatting with them at a certain time, they know whether they are supposed to be at lunch, at work, or free, and long conversations can eventually run into the character needing to leave for something else in their day. Setup involves installing Python, downloading this repository, installing its requirements, and editing a file to add your own AI provider's API key and model name before starting the server. You then point SillyTavern at the local server address it runs on, and there are a few specific settings in SillyTavern, like enabling reasoning blocks and turning off streaming responses, that the project says are required for it to work correctly. The README also describes a special character card named Setting that, if created, skips the six agent process and instead gets full access to every character's hidden internal thoughts and schedules, letting it act as a narrator, scene manager, or plot device. The author is direct that running six AI calls per message costs noticeably more in API usage than a normal single prompt setup, and explains that this cost buys a closer imitation of how human thinking actually cascades from instinct to conscious reasoning.
A Python proxy for SillyTavern that routes each character's reply through six separate AI agents simulating instinct, emotion, memory, and a daily schedule.
Mainly Python. The stack also includes Python, SillyTavern, OpenAI-compatible API.
The README excerpt provided does not state a license.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.