Pooled
A browser-based local LLM tool that pools laptop, desktop and phone GPUs for chat and coding. Open source under MIT, with no account required.
Pooled runs a single open model across browser tabs on laptops, desktops and phones, combining their memory when the model won't fit on one device. It's for people who want local AI chat or a coding assistant using hardware they already have. It's open source under the MIT license and requires no account or per-device installation.
Each device runs part of the model on its GPU through WebGPU, with direct WebRTC connections between peers. Supported models include Qwen3 1.7B, Qwen 3.8 27B and Qwen 3.6 35B MoE. Devices download only their assigned model layers, either from Hugging Face or another room member, and cache them for later use. The website provides access to the app, but inference runs on the participating devices; no server processes the conversation.
Code mode can build and edit small apps, show a sandboxed preview, inspect console errors and attempt fixes. Everyone in the room can see the files and open the preview. Files stay in the host's browser or a chosen disk folder, and edits to that folder require the host's approval.
A local API bridge connects the room to tools such as Claude Code, Codex CLI, Continue and Open WebUI. It supports OpenAI Chat Completions, OpenAI Responses and Anthropic Messages APIs, including tool calling.
Rooms share questions and answers with their members. Intermediate model data also isn't private against a determined peer, so participating devices should belong to people you trust.
