Visualize a FLUJO agent workspace as an interactive 3D network instead of flat boxes
Watch AI agent workflows execute live with animated signals showing what is running
Grow a new autonomous AI agent from a wizard using plain language questions
Pause a running agent flow and chat directly with it before resuming
| flujo-app/brain | 0xradioac7iv/tempfs | 52191314/web-agent-proxy-sdk | |
|---|---|---|---|
| Stars | 0 | 0 | 0 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 4/5 | 3/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Docker Desktop or Docker Engine with Compose, plus an existing FLUJO workspace to visualize.
brain is a visualization and management tool for FLUJO, a system for building AI agent workflows. Instead of showing agent flows as flat boxes on a canvas, brain turns a running FLUJO workspace into a 3D animated network you can fly through in your browser. Each flow becomes a glowing node, connections between flows become pulsing lines, and related flows cluster together into colored groups based on which AI provider or model they use. The visualization is live, not a static snapshot. While a flow is actually running inside FLUJO, brain listens to its event stream and animates the activity in real time: the active node lights up, connections flash as calls happen between flows, and a status strip shows what is currently executing. You can pause a running flow mid-thought and type directly to it through a chat window, with every other flow available to it as a callable tool, then resume where it left off. Beyond watching one workspace, brain includes a wizard for creating new autonomous AI agents from scratch, using three plain-language questions: where the agent's model should run, such as locally through Ollama or through a paid provider like Anthropic or OpenAI, what its goal is in one sentence, and how often it should wake up to work on that goal. Once created, each agent can learn new behaviors by writing its own new flows, install new tools from a registry, and remove behaviors or tools that are not working, all on a repeating schedule. Safety limits are built into brain's own server layer, so an agent cannot delete or overwrite its own core flow and destructive actions can require human approval. Running brain requires Git and Docker installed locally. After cloning the repository, a single docker compose up command starts both brain and FLUJO together, and the tool is reached in a browser at localhost. The project is released under the MIT license and is written in TypeScript.
A 3D live visualization tool that turns AI agent workflows built in FLUJO into an animated, explorable network you can watch think.
Mainly TypeScript. The stack also includes TypeScript, Docker, Three.js.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.