Visual AI workflows
in your browser.
NodeTool Cloud is the hosted version of the same open-source platform. No install, no GPU, no driver setup — just sign in and start building. Bring your own API keys for every cloud provider you want to use.
AGPL-3.0 source · BYOK for every provider · Self-host any time
Zero setup. Anywhere. Always current.
Cloud removes the install, the GPU, and the upgrade chores — while keeping you in control of your provider keys and your workflows.
Start in 30 seconds
Sign in, open a workflow, hit run. No installer, no native dependencies, no GPU drivers — just a browser tab.
Works on any device
Build on a laptop, tweak from a tablet, monitor from your phone. Your workflows follow you across machines.
Built for teams
Share workflows, hand off prompts, collaborate on agents. One workspace, many builders.
Bring your own keys (BYOK)
OpenAI, Anthropic, Gemini, Mistral, Groq, Replicate, FAL, ElevenLabs, HuggingFace — every cloud provider, billed to your account, not ours.
Always on the latest version
We ship Cloud updates continuously. New nodes, new providers, and bug fixes land without you lifting a finger.
No GPU required
We run the runtime, the websocket server, the queues, and the storage. Heavy generative jobs go to your chosen cloud APIs.
You pay providers directly. We never mark up tokens.
Cloud connects to OpenAI, Anthropic, Gemini, Mistral, Groq, Replicate, FAL, ElevenLabs, HuggingFace, and more — using keys you provide. No hidden inference fees, no resold credits.
What Cloud doesn't do.
If any of these matter to you, install NodeTool Studio — same workflows, same nodes, running on your hardware.
No local LLMs
Cloud doesn't run Ollama, MLX, or GGUF models — those need access to your hardware. Use Studio if you need open-weight models running locally.
Needs an internet connection
Cloud is a hosted app. If you need offline workflows, install Studio.
Your data lives with us
Workflows and assets are stored encrypted in our managed storage. If you need data to never leave your device, Studio is the right choice — or self-host the same open-source code.
Cloud is just our hosting of open-source code.
Every node, every provider, every line of the runtime that powers Cloud is on GitHub under AGPL-3.0. If you ever want to self-host, the same Docker images, CLI, and websocket server are yours to run — no vendor lock-in, no “cloud-only” features, no closed source layer.
View source on GitHub