NodeTool is the open-source creative AI workspace. Every major model from every major provider — FAL, KIE, OpenAI, Anthropic, Gemini, Replicate, and more — is called with your own keys and wired into one node-based canvas. Image, video, audio, and text live on the same surface, with editing tools like masks, inpaint, outpaint, relight, upscale, layers, and compositing built in. Runs as a desktop app on macOS, Windows, and Linux, or in the browser via NodeTool Cloud.

How is NodeTool different from ComfyUI?

ComfyUI is a Stable Diffusion power tool built for engineers. NodeTool is the full creative workspace — image, video, audio, and text on one canvas, with the editing tools creatives actually use. NodeTool also supports far more models across providers and media types, called with your own keys at provider prices.

How is NodeTool different from Figma Weave (formerly Weavy) or other closed SaaS canvases?

Closed canvases lock you into a credit system and a hand-picked list of models. NodeTool is open source. You bring your own API keys to every provider, pay providers directly at provider prices, and own your workflows and files. Cloud is just our managed hosting of the same open-source code you can run yourself.

How does pricing work?

NodeTool Studio is free to download and use. NodeTool Cloud is a subscription for managed hosting. In both editions, you bring your own API keys to every provider and pay those providers directly at their list prices. NodeTool does not run models for you on its own servers, does not issue its own credits, and does not mark up model calls.

What models does NodeTool support?

The latest models, including Flux, Seedance, Wan, Veo, Kling, Hailuo, Qwen Image, Whisper, ElevenLabs, and Suno, called through providers like FAL, KIE, OpenAI, Anthropic, Gemini, Replicate, Together, Groq, Mistral, OpenRouter, and HuggingFace. Models can also run on your own machine via MLX, Ollama, llama.cpp, vLLM, and LM Studio.

Independent generative artists, motion designers, AI-native illustrators, technical art directors, ComfyUI power users frustrated with the UX, and small creative studios, brand teams, and post-production shops working with AI every day.

Is NodeTool open source?

Yes. Both Studio and Cloud share the same AGPL-3.0 codebase. There is no closed-source layer and no "pro tier" hiding the good features. You can self-host any time.

All templates

Template·Agents & Research

Model Arena

Name: NodeTool
Author: NodeTool

One brief, three frontier models, answered side by side. The same prompt fans out to OpenAI, Anthropic, and Google in a single pass, each answering in the same structure — so you compare reasoning, not formatting. Swap the model on any lane to build your own bracket.

Download NodeTool How to run it Try it as a mini app

Model Arena — example output from the NodeTool workflow

The workflow

Workflow EditorModel Arena

Note

String Input

Should a 12-person B2B SaaS startup build its own customer-facing analytics dashboard, or integrate a third-party embedded-analytics vendor…

String Input

Audience: the founding team (technical). Constraints: 2 engineers can be spared for 6 weeks; weak reporting is a top-3 reason prospects chu…

Prompt

Brief

Context

Brief: {{ BRIEF }} Audience & constraints: {{ CONTEXT }} Answer in Markdown with exactly these three sections and nothing else: ## Recommen…

Agent

Prompt

Text

gpt-5-mini

Agent

Prompt

Text

claude-sonnet-5

Agent

Prompt

Text

gemini-3.5-flash

Output

Nodes in this workflow

9 nodes · 4 types

Agent
nodetool.agents.Agent
×3
Output
nodetool.output.Output
×3
String Input
nodetool.input.StringInput
×2
Prompt
nodetool.text.Prompt

How to run it

01
Download NodeTool Studio
Install the free desktop app for macOS, Windows, or Linux. It runs on your own machine, no account required to start.
02
Open the Model Arena template
Browse the built-in template library inside Studio and open this workflow onto the canvas. Every node is already wired up.
03
Add your keys
Connect the providers this workflow uses (Agent). Bring your own keys — you pay the provider directly.
04
Run and remix
Hit Run to execute the graph and watch results stream in. Swap models, edit prompts, or rewire nodes to make it yours.