vs Image-only node tools (like ComfyUI)
Specialized node tools only create pictures. NodeTool provides a complete studio that connects images, video, audio, text, and editing on one canvas.
Free & Open Source
You direct the vision.The agent builds the film.
Describe your idea. The agent writes the script, boards every scene, generates the footage, and cuts the timeline — end-to-end production on an open node-based canvas, with complete creative control.
- Script, storyboard, sketch, timeline, 3D
- Your own API keys, no token markups
- Open source, you own the files
The way it works today
One film
shouldn't take five separate apps.
Direct it on one canvas. Run it on your own accounts. Keep the editable project when you're done.
See how NodeTool compares- Five browser tabs per shot
One app for images, another for animation, a third for voiceovers, and a separate editor to stitch it together.
- Export, import, repeat
Every hop between apps drops your settings, your prompts, and the reasoning behind them.
- Token markups on every take
Credit packs priced above what the models cost, spendable only on the models that platform picked.
- One flat file, no way back in
A finished render you cannot re-cut, in a project format that opens nowhere else.
A chatbot gives you a transcript.
The agent gives you the project.
NodeTool's agent operates the studio for you. It lays the steps out on your canvas, generates the takes, and leaves every part on the table — so you can swap a model, re-roll a shot, or run the whole thing again.
How agents and workflows work together →
It reads the studio before it answers
The agent checks what this install actually has — nodes, packs, example workflows — instead of guessing from a system prompt. The answer is a map of your workspace, not of the product.
Every step runs in the open
A storyboard created, six keyframes rendered, a contact sheet built — each call is a card you can expand to see the code it ran and what came back. No spinner standing in for the work.
You see the bill before you pay it
Your keys, your invoice. Asked to price the render, the agent reads the providers' own pages and puts the four candidates side by side, so the decision is yours and not a credit balance's.
It leaves documents, not a transcript
The teaser is cut, and so are a storyboard, a timeline and a reusable workflow — each one a link that opens the editor. Change a shot and re-run; nothing has to be rebuilt from the chat.
A trailer generated end-to-end in NodeTool — script, key art, animation, and sound from one workflow on the canvas.
01Pitch
Tell the agent your concept, style, and tone. It drafts the script, casts the voices, and sets the visual direction. Prefer to wire it yourself? It's the same canvas.
Audio Input
Whisper
GPT Image 2
02Automate
The agent storyboards scenes, generates the footage, syncs the audio, and cuts it all together on a multi-track timeline — running on your own accounts at provider prices.
Logs2.4s
- Workflow started
- Text written
- Image generated
- Clip assembled
- Workflow finished
Preview
1024×1024PNG
03Direct
Jump in at any moment. Swap a voice take, re-roll a clip, or trim frames. The agent builds the foundation; you refine the final cut.
Editors
Use cases
From a pitch to a finished cut
Real projects you can open, run, and re-cut. Each one runs from the pitch to the finished piece without leaving the canvas, and every one is built from the blocks you get on day one.
Film
Movie Trailer Generator
Type one logline and get a storyboard back: a card per shot, a still on every card. Approve the ones you like, animate those, and cut the trailer on the timeline.
LoglineStoryboardStillsClipsTimeline
View use caseDocumentary
Documentary Teaser Generator
Describe the film in a sentence and the board comes back readable: a card per shot, a still on every card. Approve the stills, animate them, and cut the teaser on the timeline.
PremiseStoryboardStillsClipsTimeline
View use caseMarketing
Product Video Generator
Turn a campaign brief and a single product photo into a cinematic 16:9 product video. Your inputs feed a prompt, an agent directs the shot, and a text-to-video model renders it.
BriefPromptAgentText-to-Video
View use case
Design
Movie Poster Generator
From a title, genre, and audience, the canvas writes a creative strategy and renders a batch of cinematic movie poster concepts, title, tagline, billing block and all.
BriefStrategyConceptsKey art
View use caseProfessional post, right next to the models
Cut and refine on the same canvas you generate on, with no export hops. A storyboard becomes a script, the script becomes takes, the takes land on a multi-track timeline — and the agent works every one of them through the tools you click.
Visual storyboards
Board the film shot by shot. Generate cheap stills first to lock the look, then pay to animate only the shots you approved.
More on storyboard →Scripting & casting
Draft the dialogue, cast a voice per character, and audition alternate line readings. Change the words and the take flags itself stale, so you see exactly what still needs voicing.
More on script & voice →Multi-track timeline
Arrange, trim, and layer generated video and audio across multiple tracks, down to the frame and the stem. The agent edits the same document when you ask it to tighten the opening.
More on timeline →Layered drawing canvas
Sketch, paint with brushes, and blend hand-drawn elements with AI-generated layers. Bind a layer to a prompt and regenerate that layer alone.
More on sketch →Spatial composition you can reproduce
Place primitives and lights in a glTF scene by hand or by tool call. The same operations run headlessly, so a scene is reproducible.
One workspace, several doors
Choose the setup that fits your workflow.
Whether you use the desktop app or the browser version, you get the exact same features, tools, and capabilities: the same agent, the same workflows, the same models.
Desktop app
Studio
Runs directly on your computer. Keeps your files and settings private on your hard drive, with optional offline support.
Go to StudioIn your browser
Cloud
Runs in your web browser with zero installation needed. The fastest way to jump in and start creating. Currently in alpha.
Go to CloudFor creators
Creatives
Focus on the creative vision — images, video, voice — and let the software handle the technical plumbing.
Go to CreativesFor developers
Developers
Connect custom scripts, automate tasks, and integrate your own code into the visual canvas, through the SDK, the CLI, or an MCP server.
Go to DevelopersProduce campaigns
Marketing
Turn a brief into a production workflow you run again for every product, market, and variant.
Go to MarketingComparison
Bring all your creative tools under one roof.
You're probably using two or three of these already. Here's what changes when they all live on one canvas.
vs Closed platforms
Other platforms force you into monthly credit bundles and restricted feature lists. NodeTool is open source: you use your direct provider accounts and pay raw rates with no middleman markup.
vs A dozen separate web apps
Instead of copying files between disconnected websites, NodeTool connects your favorite AI tools into a single, repeatable workflow.
How does NodeTool compare with ComfyUI and Figma Weave?
ComfyUI goes deep on diffusion images and runs locally. Figma Weave is a hosted canvas with a curated model list, billed in its own AI credits. NodeTool is open source, covers image, video, audio, and text, and calls every provider on your own keys.
| NodeTool | ComfyUI | Figma Weave | |
|---|---|---|---|
| Media on one canvas | Image, video, audio, text | Diffusion images | Hosted creative canvas |
| Models | Every major provider and media type | Stable Diffusion and other diffusion models | Hand-picked list |
| What you pay | Provider list prices, on your own keys | Your own compute and provider accounts | Figma Weave AI credits |
| Source | Open source (AGPL-3.0) | Open source | Closed |
| Where it runs | Desktop app and browser, self-host any time | Your own machine | Browser only |
Where NodeTool fits
Model freedom. No platform lock-in.
Take Seedance, one of today's best video models. It is sold by FAL, Replicate, and KIE at different prices, and NodeTool lets you pick the cheapest of the three. When the next Veo or Kling arrives, you switch to it in one click.
That is what sovereignty buys you: the best model at the best price each week, and nothing to lose if your favorite tool gets bought, repriced, or shut down.
Local. Open. Yours.
No middlemen, no markups.
No locked project formats, no token markups, zero subscription traps. Run inference locally or bring your own API keys, and switch between providers with one click. You own the files, the workflow, and the final cut.
Bring your own keys
Connect directly to OpenAI, Anthropic, Gemini, FAL, KIE, Replicate, and the specialized video models. Your keys stay on your disk in Studio, encrypted in Cloud.
No token markups
No credit packs, no subscription traps. If an image costs $0.03 at the provider, you pay $0.03 to the provider. NodeTool takes no cut.
Open source, end to end
Studio and Cloud are built from the same AGPL-3.0 source, with no paywalled features. Read it, fork it, or host it yourself at any time.
Run it on your own hardware
Download open-weight models and run inference locally — MLX, Ollama, llama.cpp, vLLM, and LM Studio — with no internet connection.
Model freedom. No platform lock-in.
Route your shots through the best video, image, audio, and language models on the market — or run open weights locally on your own hardware. Switch between them with one click.
Latest Models
GPT-5.6Claude Fable 5Claude Opus 5Claude Sonnet 5Gemini 3.5 FlashGemini 3.1 ProQwen ImageVeo 3.1Kling 3Seedance 3Hailuo 2.3Wan 2.5FLUXWhisperElevenLabs
Cloud Providers
Cost transparency
See what every run actually costs
NodeTool records what each step costs and adds it up by workflow, provider, or model. Because you bring your own keys and pay providers directly, with no token markups, these are the amounts you are actually billed.
Costs
Costs
Spend per node execution across 9 workflows · last 90 days
7d14d30d90d
Export CSVTotal spend
$7.71
across 9 workflows
Node executions
194
0 failed
Avg / execution
$0.040
across all node types
Top cost driver
BytedanceSeedance2
$3.08 · 40% of spend
Spend over timedaily · last 90 days
Kie$4.35OpenAI$1.85fal.ai$1.50
$5.39$2.70$0
Group by
ExecutionNode typeWorkflowProviderModel
5 of 31Node typeRunsShare of spendCost
BytedanceSeedance21$3.08
40%
GptImage2Edit10$1.41
18%
gpt-5.4-mini37$1.25
16%
GrokImagineTextToVideo21$1.01
13%
gpt-5-mini29$0.34
4%
Cost per step
Every run is priced on its own and can be grouped by step, workflow, provider, or model.
No markup
Pay OpenAI, fal.ai, and the rest directly with your own keys. NodeTool takes no cut.
Export anytime
Download the full breakdown as a spreadsheet for your own books or client invoices.
The same model, priced twice
NodeTool bills you nothing per generation — you hold the keys and pay each provider its list price. Set a workload and compare it with the same models bought as Figma Weave credits.
- 750 images / month$0.035 / image
Stills, variations, and every reroll you throw away.
- 24 clips / month$0.0175 / second
A provider bills the second, a credit plan bills the clip — so the length matters to one side and not the other.
- 120 minutes / month$50 / 1M characters
Narration minutes, billed by character (~950 per minute). Figma Weave has no voice model, so this line has nothing to compare against.
Your keys, direct
$35.31/ month
- Images · Seedream V5$26.25
- Video · Seedance 1.5$3.36
- Voice · Eleven Turbo v2.5$5.70
- NodeTool Studio$0
The same models on Figma Weave
$45.00/ month
3,744 of 4,000 credits. Against $29.61 for the same images and clips on your own keys.
4 credits an image, 31.01 a clip, derived from Figma Weave's own table. It prices a clip, not a second, and does not publish clip length.
The picker holds the 10 models Figma Weave sells that Kie or AtlasCloud also serves, so both columns price the same model. Its other 13 are left out because neither provider carries them. Prices via genspend.io, updated 2026-08-23. Figma Weave plans read 2026-08-24. Local models through Ollama, MLX, and llama.cpp cost nothing per call.
One canvas.
The whole production.
Script, storyboard, footage, sound, and the cut — on one canvas, with the editing tools sitting right next to the models. You direct the whole piece instead of generating parts of it.
Workflow Editor
Scripting & casting
Go from prompt to screenplay in seconds. Draft dialogue, cast character voices, and audition alternate line readings with automatic word-level sync.
Any model, local or cloud
Route your shots through Flux, Seedance, Wan, Whisper, ElevenLabs, Suno, and the rest — or run open weights locally on your own hardware. No platform lock-in.
A real timeline, not a black box
Instead of a single unfixable video file, the agent hands you a fully editable multi-track project. Fine-tune every frame, layer, and audio stem.
Edit where you generate
Mask, retouch, extend, relight, upscale, layer, and composite. The tools you reach for live on the same canvas as the models.
Storyboard before you spend
Pitch a concept and a visual style, press Direct, and get a shot list back — action, camera, motion, and duration per card. Approve the stills you like, animate only those, then lay the whole thing onto the multi-track timeline in one click.
Stills cost cents, clips cost dollars
Generate stills until one looks right and pick it. Only then do you pay to animate the shot into a clip.
Revise one shot, not the reel
"Darker, add rain" runs video-to-video on that clip and swaps it in place. Fixing shot 3 never re-rolls shots 1–5.
A cast that stays consistent
Characters, locations, styles, and props are named entities with a fixed descriptor that goes into every prompt that names them.
Scripting & casting
Go from prompt to screenplay in seconds. Draft the dialogue, cast a voice per character, and audition alternate line readings with automatic word-level sync. The script is its own document — the words are the source of truth, and the takes are derived from them.
A voice per speaker
Cast each speaker once — provider, model, voice — and every line they own inherits it until you override that line.
Takes you can audition
Voicing a line saves a take with its own word timings. Keep several, compare them, pick the read you want.
Edits mark themselves stale
Change a line's words or its voice and the line flags itself against the take that no longer matches. Re-voice that line, not the script.
A real timeline. Not a black box.
Instead of a single unfixable video file, the agent hands you a fully editable multi-track project. Generate at the playhead, then fine-tune every frame, layer, and audio stem before you export.
Generate at the playhead
Prompt a model and the new clip lands on the track right under the playhead — no import step.
Multi-track video & audio
AI video and AI audio share one timeline. Stack tracks, trim, split, and rearrange clips.
Export a finished cut
Render the whole sequence to one video. The cut happens where you generate, not in another app.
Built-in sketch editor
A full layer-based image editor with generation built in. Paint with real brushes, stack layers and blend modes, then prompt a model to create or refine any layer — and export a finished PNG without leaving the canvas.
A real brush engine
Round, soft, airbrush, and spray brushes with size, opacity, hardness, and angle. Pick any color and paint straight onto the canvas.
Layers & blend modes
Stack layers, set opacity and blend modes, reorder and mask — the same controls you reach for in a pro image editor.
Generate onto a layer
Describe an image and it lands on its own layer. Mix hand-painted strokes with text-to-image and image-to-image right where you draw.
Wrap the graph in a screen
A workflow is for you. An app is for everyone else — the same graph with inputs, a run button, and a place for the result. Here is Image Enhance, and the app it became.
Nothing is regenerated in between. The app names the workflow's inputs and outputs, so a change to the graph shows up in the screen — and an agent can place the widgets for you. Open Photo Studio →
Turn workflows into simple apps you can share
Once you build a workflow, turn it into a clean interface with a few input boxes and a Generate button. Hand that screen to a teammate or client so they get results without navigating the canvas underneath — which stays yours to open and change.
Drag a widget, wire a port
Fields, sliders, pickers, and result displays bind to the inputs and outputs of the workflows the app runs. Results stream in as the run happens.
An agent builds and grades it
Describe the app and it plans the workflows, places the widgets, replays every interaction, and has a second model judge whether the result matches the ask. No passing verdict, no app.
One file to hand over
Export an app as a bundle — the screen plus the full graph of every workflow behind it — and import it on another machine.
All the
building blocks
Hundreds of ready-made blocks for models, data, and files, plus one for every model on Replicate, FAL, and Kie.ai.
Organize Everything
Built-in Asset Manager
Drag files in once, use them in any workflow.
Import & organize
Drag and drop. Files sort themselves by type, project, or tag.
Preview files
Built-in preview for images, audio, video, and documents.
Use in workflows
Point any workflow at a single file or a whole folder.
Images & Graphics
PNG, JPG, GIF, SVG, WebP
Audio & Video
MP3, WAV, MP4, MOV, AVI
Documents & Data
PDF, TXT, JSON, CSV, DOCX
Hello
HuggingFace
Search, download, and manage models straight from Hugging Face.
Organize your local library without leaving the app.
Model Manager
Frequently asked questions
What is NodeTool?
NodeTool is the open AI production studio. You direct the vision and the agent builds the film: it drafts the script, boards every scene, generates the footage, and cuts a multi-track timeline you can still edit. Every major model from every major provider, including FAL, KIE, OpenAI, Anthropic, Gemini, and Replicate, sits on one visual canvas that you run on your own machine or in the browser.
Read more →Is NodeTool open source?
Yes. The whole project is AGPL-3.0 on GitHub. Studio and Cloud are built from the same source, and nothing is held back for a paid tier. You can host it yourself at any time.
Read more →How much does NodeTool cost?
NodeTool Studio, the desktop app, is free and open source. NodeTool Cloud is a subscription that covers hosting; it is in alpha and its price is set at full release. In both editions you bring your own API keys and pay each provider their list price directly, so what a run costs is whatever the provider charges for it. Local models through Ollama, MLX, or llama.cpp cost nothing per call.
Read more →Is NodeTool local-first and private?
NodeTool Studio runs on your own machine: the workspace, your workflows, your files, and your API keys stay there, and open-weight models can run locally through MLX, Ollama, llama.cpp, vLLM, and LM Studio. When you call a hosted model, that request goes from your machine straight to the provider you chose, on your key — it does not pass through NodeTool servers. Cloud is the same open-source app hosted in the browser, and your keys still call the providers directly.
Read more →Which models are supported?
The leading models, including Flux, Seedance, Wan, Veo, Kling, Hailuo, Qwen Image, Whisper, ElevenLabs, and Suno, reached through providers such as FAL, KIE, OpenAI, Anthropic, Gemini, Replicate, Together, Groq, Mistral, OpenRouter, and HuggingFace. Models can also run on your own computer through MLX, Ollama, llama.cpp, vLLM, and LM Studio.
Read more →What does it mean that NodeTool is agent-first?
Every editor in NodeTool — the node canvas, sketch pad, storyboard, video timeline, script editor, 3D scene, and app builder — is exposed to agents as tools, around 120 in all. An agent doesn't describe what you could do; it builds the workflow, runs it, and repairs what fails, on the same surfaces you use. The same tools are exposed over MCP, so outside agents such as Claude Desktop and Claude Code can drive NodeTool too.
Read more →Do I need a GPU to run NodeTool?
No. NodeTool works entirely with cloud providers — bring your own API key, and the model runs on the provider's hardware with nothing to install locally. A GPU only matters if you choose to run open-weight models yourself through Ollama, MLX, or llama.cpp: it speeds those up, but many smaller local models run fine on CPU too, just slower.
Read more →Made 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.
Ready to call action?
Download the open-source studio for macOS, Windows, and Linux. Free, AGPL-3.0, and built alongside working filmmakers, designers, and video editors. Or open it in your browser with nothing to install.