Build Multiplayer AI Bots Locally in Minutes with OpenBot 3

Description and use

OpenBot 3 enables rapid local deployment of deterministic multiplayer AI bots without cloud reliance, leveraging quantised LLMs and WebRTC.

Problem

Developers want to run fully autonomous, multiplayer‑ready AI bots on a local machine without cloud subscriptions, data leaks, or high costs.

Outcome

OpenBot 3 enables rapid, local deployment of deterministic multiplayer AI bots, eliminating cloud costs and ensuring data sovereignty.

Workflow

• Clone the OpenBot 3 repository and install Rust toolchain
• Download a quantised LLM (e.g., LLaMA‑7B‑GGUF)
• Build the server binary with cargo
• Launch the server specifying model and port
• Run client instances pointing to the local server via WebRTC
• (Optional) Register plugins and launch with plugin flag
• (Optional) Connect clients in other languages such as JavaScript via WASM

Evidence

Clone the repo and install the Rust toolchain

Source

Pull a quantised LLM (LLaMA‑7B‑GGUF ≈ 4 GB)

Source

Build the core and launch a local server

Source

Run two client instances (they’ll connect via WebRTC)

Source

All three projects report 0 % cloud spend, full data sovereignty, and iteration cycles under 30 seconds

Source