read_the_room / README.md
Ilia-Iliev's picture
Upload folder using huggingface_hub
ad4c37d verified
|
Raw
History Blame Contribute Delete
2.4 kB

A newer version of the Gradio SDK is available: 6.22.0

Upgrade
metadata
title: Read The Room
emoji: ๐Ÿ‘€
colorFrom: yellow
colorTo: yellow
sdk: gradio
sdk_version: 6.16.0
python_version: '3.13'
app_file: app.py
pinned: false
tags:
  - track:wood
  - sponsor:openai
  - sponsor:modal
  - achievement:offgrid
  - achievement:offbrand
  - achievement:llama
  - achievement:sharing
  - achievement:fieldnotes

Track 2: Read the Room

Play here

You're a player in a situation where each character has an agenda and an opinion of everyone else, including you. You talk your way to a goal. The inspiration is text adventures RPGs like Zork crossed with social deduction games like Avalon and Werewolf. The social dynamic is the game - I've always wished videogame dialogues gave me more options to explore. Traditional games have always been constrained in this and current AI games dodge the social element: they usually focus on one character, or fixed characters on rails. So I built a game with multiple AIs reacting to each other and especially you (the player)

Anyone can publish a scenario (optional but recommended scene picture) and play what others shared

Merit Badges

Off the grid

The app makes zero network requests of its own - art and font are bundled and inlined. Point RTR_API_BASE at a local llama.cpp/vLLM and it runs fully offline (the hosted Space points it at a Modal GPU instead)

Off-Brand

The default Gradio doesn't work well for a game, so I gave it some style with a custom front-end

Llama champion

Llama.cpp is the canon way to play the game - start a server locally and wire it up. I playtested with Qwen3.6-27B (Q8_K_XL gguf from unsloth)

Sharing is Caring

Playthrough traces - raw prompt and completions are available. You can use ๐Ÿž export function to see what happened each turn - what fed into the LLM and what the response was. Here is one: Ilia-Iliev/read-the-room-traces

Field Notes

More details in my blogpost

Video

Social Media post