Skip to content

NodeTool Cloud is in alpha. Try it →

All solutions
For local-first builders

Local-First AI Workflows

Run a workflow on an open model on your own computer. Below is one run on Ollama, the memory each model needs, and what still works when you disconnect.

  • Ollama, MLX, llama.cpp, vLLM, or LM Studio
  • No account, and no API key for local models
  • Update checks and analytics stay off until you turn them on
  • Hosted models only on the nodes where you pick one

The workflow behind it

Data flow

One local run

Ask your notes a question. The model runs on your machine.

The Private Assistant template answers from the text you paste in and nothing else. It is built for a local Ollama model, so the answer comes from the computer you are sitting at.

  1. 01

    Pull a model

    Install Ollama and pull a chat model, or download one from the Model Manager in Studio. This is the only step that needs the network.

  2. 02

    Pick it on the Agent node

    Open the Private Assistant template and choose the Ollama model on its Agent node. The graph has no API key field to fill in.

  3. 03

    Run

    The Prompt node joins the notes, the question, and the tone. The Agent node sends them to Ollama on 127.0.0.1:11434, and the answer comes back to the Output node.

Private Assistant · inputsollama · 127.0.0.1:11434
document
Team sync, Thursday - Launch date moved to March 14 (was March 7) to finish the accessibility fixes. - Budget: $8k left for Q1, reserved for the usability study. - Open risk: the billing migration hasn't been tested on staging yet.
question
When do we launch, and what's the biggest open risk?
tone
concise and neutral

The inputs as they ship with the template, shortened. Both answers are in the notes: the March 14 launch and the untested billing migration. Paste your own notes to replace them.

What it needs

Pick the model by the memory you have.

Memory means graphics memory, or unified memory on an Apple Silicon Mac. Ollama also runs on the CPU, more slowly. The Model Manager in Studio compares these figures with your machine before you download.

Local chat models for the Private Assistant, with download size and memory
Model on OllamaDownloadMemoryNote
Qwen3.5 2B2.7 GB4 GBRuns without a GPU
Qwen3.5 9B6.6 GB8 GBChat and tool use
GPT-OSS 20B14 GB16 GBNodeTool's default local chat model
Qwen3.6 27B17 GB20 GBThe largest of the four

Approximate figures from the Model Manager. GPU requirements for image, speech, and video models.

What stays on the machine

Unplug the network and run it again.

Your text
The notes, the question, and the answer go from Studio to Ollama on the same machine. No request leaves it.
The project
The workflow, its inputs, and the run log are stored by Studio on your disk.
Account
None. Studio does not ask you to sign in to run a local model.
Background calls
Update checks and product analytics are off until you turn them on in Settings.

With the network off

Local nodes run. Hosted nodes stop where they are.

Still works
This workflow, the editors, your project files, and any other node set to a model you already downloaded.
Stops at the node
A node set to a hosted model fails when it runs, and the run log names that node. Swap it for a local model and run again.
Needs a connection
Downloading models, web search and fetch nodes, and hosted providers.

Frequently asked questions

Can one workflow mix local and hosted models?

Yes. Each node has its own model. Keep the private step on a local model and send only the step that needs a hosted model to that provider. Only that node needs the network.

Can it run image, speech, and video models locally too?

Yes. Image and speech models such as Stable Diffusion XL, FLUX.1 schnell, Whisper, and Kokoro run through the HuggingFace pack on an NVIDIA card, an AMD card on Linux, or an Apple Silicon Mac. Local video generation needs an NVIDIA or AMD card. The Model Manager lists the models that fit your machine.

Does Cloud run local models?

No. Cloud runs on our servers and uses hosted providers. Local models need Studio, the desktop app.

Build it on your own machine

Download the open-source studio for macOS, Windows, and Linux. Open the template and direct it yourself, on your own keys.