Build AI workflows that
run on your machine.
NodeTool Studio is the desktop app for people who want their models, data, and results to stay on their own machine. Run free local models with Ollama and MLX, and bring your own keys for a cloud provider whenever you need one.
Free · AGPL-3.0 · macOS, Windows, Linux · No account required · All downloads
Privacy, speed, and no usage fees.
Everything Studio does, it can do without a network connection. Your prompts, your files, and your models all stay on your own hardware.
Works fully offline
Once your local models are downloaded you can disconnect from the internet and keep working, whether you are travelling or on a machine with no network at all.
Your data never leaves the device
Workflows, files, prompts, and results stay on your disk. Nothing is reported back to us, and nothing leaves the machine unless you call a remote provider on purpose.
Run open models on your machine
Ollama, MLX for Apple Silicon, llama.cpp, and Hugging Face all work with the same building blocks. Pick any open model and pay no usage fees.
Use your GPU to the fullest
NVIDIA graphics cards on Windows and Linux, Apple Silicon on macOS. Image, video, and audio work runs on your own hardware.
Own your model library
The built-in model manager downloads and organizes model files and shares them across workflows, so you keep exactly the models you want.
Custom Python and TypeScript nodes
Add your own code, install Python packages in the environment that ships with the app, or run code you do not trust inside Docker.
What Studio asks of you.
Running everything on your own machine means taking on a few jobs the cloud would otherwise handle. If any of these feel like too much, NodeTool Cloud runs the same workflows in your browser.
Hardware matters
Local models need memory and ideally a graphics card. For serious local work we suggest 16GB or more of RAM and at least 4GB of graphics memory.
You manage updates
When a new release comes out, you install it. Builds are signed and notarized for macOS and Windows, so updating stays simple.
Disk space
Local models are large. Allow around 20GB for a small starter set, and considerably more for image and video models.
The newest models, the day they arrive.
The leading image, video, audio, and language models from every major provider, or the same kind of models running on your own computer when you prefer that.
Hello
HuggingFace
Search, download, and manage models straight from Hugging Face.
Organize your local library without leaving the app.
One canvas
for the whole craft
Image, video, audio, and text on a single visual canvas, with the editing tools you already rely on sitting right next to the models. You direct the whole piece instead of generating parts of it.
Edit where you generate
Mask, retouch, extend, relight, upscale, layer, and composite. The editing tools you reach for live on the same canvas as the models.
Watch every step render
Results appear as each step finishes. Inspect any frame, swap a model, and re-run from that point on.
Your keys, provider prices
Bring your own keys for FAL, KIE, OpenAI, Anthropic, Gemini, Replicate, and the rest. The bill comes from the provider, not from us.
Image, video, audio, text
Flux, Seedance, Wan, ControlNet, Whisper, ElevenLabs, and Suno, all on one canvas under their real names. You always know which model you are running.
Studio runs on your machine.
Cloud runs in your browser.
Same workflows, same building blocks, same providers. Pick the one that suits how you work, and switch whenever you like. Both are AGPL-3.0 open source: Cloud is simply our hosting of the same code you can run yourself.
NodeTool StudioYou are here
Local-first desktop app
Best for: artists with a good graphics card or an Apple Silicon Mac, big local model collections, offline work, and anyone who wants everything kept on their own disk.
- Where it runsYour machine (macOS, Windows, Linux)
- Install requiredDesktop app + ~20GB for local models
- Local models (Ollama, MLX, GGUF)Yes — runs entirely on your hardware
- Bring your own API keysAll providers — keys stored locally
- Works offlineYes — fully offline with local models
- Where your data livesOn your disk only
- GPU requirementsRecommended for local models
- UpdatesYou install new releases
- Source code100% open source (AGPL-3.0)
- CostFree — pay only for the cloud APIs you use
NodeTool CloudAlpha
Hosted in the browser · Alpha preview
Best for: studios and solo artists who want to start in seconds, work from any device, and skip the hardware setup, while still using their own keys with every provider.
- Where it runsHosted by us — open in any browser
- Install requiredNone — sign in and start building
- Local models (Ollama, MLX, GGUF)Not available — cloud APIs only
- Bring your own API keysAll providers — keys stored encrypted
- Works offlineNo — needs an internet connection
- Where your data livesOur managed storage (encrypted at rest)
- GPU requirementsNone — heavy jobs run on cloud APIs
- UpdatesAlways on the latest version
- Source code100% open source (AGPL-3.0) — self-host any time
- CostFree during alpha — you pay only your own API costs
Not sure? Start with Cloud to try it in 30 seconds, then move to Studio when you want everything on your own machine. Your workflows move freely between the two.
100% open source. Always.
Studio is released under AGPL-3.0. Every building block, every provider, and every line that runs it is on GitHub. Read it, copy it, or host it yourself. There is no separate paid version.
View source on GitHubMade with working creatives
NodeTool is open source under AGPL-3.0. Star the project on GitHub, join us on Discord, and share workflows with the artists, motion designers, and studios already using it on real jobs.
Get in
touch
Questions, bug reports, feature requests — we answer.