GriptapeIn this video we’ll show you the new Library Management feature in Griptape Nodes, our visual editor for creative professionals who want to bring generative AI into their image, video, and audio workflows.
Until now, adding new node libraries required using the command line and editing engine settings manually. With this update, you can now manage and install new libraries directly from within Griptape Nodes, no Git commands required.
🔧 What You’ll Learn: - How to access the new Library Management panel - Installing community and partner libraries from GitHub with one click - Setting up API keys and secrets for libraries like Black Forest Labs - Refreshing libraries and using new nodes instantly inside your canvas
This feature makes extending Griptape Nodes faster, easier, and more accessible than ever, letting you focus on creating instead of configuring.
Easily Manage and Install Libraries in Griptape NodesGriptape2025-12-09 | In this video we’ll show you the new Library Management feature in Griptape Nodes, our visual editor for creative professionals who want to bring generative AI into their image, video, and audio workflows.
Until now, adding new node libraries required using the command line and editing engine settings manually. With this update, you can now manage and install new libraries directly from within Griptape Nodes, no Git commands required.
🔧 What You’ll Learn: - How to access the new Library Management panel - Installing community and partner libraries from GitHub with one click - Setting up API keys and secrets for libraries like Black Forest Labs - Refreshing libraries and using new nodes instantly inside your canvas
This feature makes extending Griptape Nodes faster, easier, and more accessible than ever, letting you focus on creating instead of configuring.
Griptape Nodes enables you to integrate AI models or agents directly into your VFX workflow, giving you secure and granular control over your creativity.
To install Griptape Nodes: 1 - Head to griptapenodes.com 2 - Click Download 3 - Choose the download for your relevant OS 4 - Run the installer 5 - Create an account or sign in 6 - Add a workspace directory 7 - Start the engine 8 - Launch the Editor and start building AI workflows!
Join our Discord channel: discord.gg/griptapeFacial enhancement workflow in Foundry Griptape NodesGriptape2026-03-04 | Here's a complete, node-by-node walkthrough of a face enhancement workflow using Foundry Griptape Nodes. This video demonstrates how to detect faces in an image, improve their detail and quality using an open-source diffusion model, and then merge the enhanced faces back into the original image for a final composite output.
Generative Models used in this workflow: - Sample image Generation: Google Nano Banana Pro API node from the Griptape Nodes Standard Library - Face Detection: Ultralytics YOLOv8 Face Detection from the Griptape Nodes Advanced Media Library) - Face Segmentation: Meta AI SAM3 from the SAM3 Add On Library for Griptape Nodes (github.com/griptape-ai/griptape-nodes-library-sam3) - Face Detail Enhancement: Diffusion Pipeline Builder Node, with a custom pipeline using the FLUX-Krea-BLAZE (huggingface.co/MintLab/FLUX-Krea-BLAZE) model, coupled with the Generate Image Diffusion Pipeline Node
Note that the Krea BLAZE repo on hugging face only includes the .safetensors file for the model checkpoint. To use Krea BLAZE you must copy the other assets from the FLUX.1-Krea-dev repo at huggingface.co/black-forest-labs/FLUX.1-Krea-dev/tree/main into the directory in your Hugging Face cache where the Krea BLAZE .safetensors file is located.
The simplest way to do this is to pull the FLUX.1-Krea-dev using the Hugging Face CLI on the machine where your Griptape Nodes engine is running or the Griptape Nodes model manager and then copy the files using the CLI or file explorer on your engine machine.
You will see how to create a node that transfers the color palette from one image to another. In the example, a gritty oil refinery palette is applied to a wheat field, producing a cohesive, post-industrial look. This kind of color consistency is especially useful when preparing assets for image-to-video workflows, animation, or multi-shot sequences.
Ian breaks down the full development process step by step, including how to set up a clean node library, how to use the Griptape Nodes library template, and how to leverage the Node Development Guide as structured context for an AI coding assistant. He shows exactly how to prompt the agent with requirements for UI controls, parameters, outputs, and third-party color transfer algorithms.
After generating the code, the new node is loaded into Griptape Nodes and tested against the original implementation. The result is a perfect match, demonstrating a successful one-shot generation.
This video is a practical guide for anyone looking to extend Griptape Nodes with custom tools, whether you are a creative technologist, developer, or artist seeking more control over your workflow.
=== Complete prompt:
Looking at the guidance in @README.md and the node development guide referenced in that file
Create new ColorMatchDemo node for Griptape nodes
The ColorMatchDemo node transfers color characteristics from a source image to a target image. High-quality and consistent color matching is achvieved using advanced color transfer methods, including those proposed by Reinhard et al., Pitie et al., and a Multi-Variate Gaussian Distribution (MVGD) transfer combined with classical histogram matching. The node's function is to enable artists to perform efficient color harmonization across different images, improving the visual coherence their projects.
Node parameters (in order of appearance in the node UI)
- Source Image: The image from which the color characteristics will be extracted (accepts a ImageUrlArtifact or ImageArtifact as input) - Target Image: The image to which the color characteristics will be transferred (accepts a ImageUrlArtifact or ImageArtifact as input) - Color Transfer Method: The method to use for the color transfer (drop down menu with options mkl (default), hm, reinhard, mvgd, hm-mvgd-hm, and hm-mkl-hm) - Color Transfer Strength: The strength of the color transfer (float slider with range 0-10 (Default 1). With the stength set to 1, color should be fully transferred) - Output Image: The image with the transferred color characteristics (output is an ImageUrlArtifact)
Create the new ColorMatchDemo Node in this directory:
Before using the model you will need to accept the gated terms on Hugging Face and download the SAM 3 files. You will also need a Hugging Face token inside your Griptape Nodes settings. SAM 3 currently runs on Windows with CUDA, so your engine should run on a Windows machine.
What you will learn: • How to install the SAM 3 extension through the new Library Management interface • How to run SAM 3 on images and interpret the composite and mask outputs • How to use masks for inpainting and background replacement with models like Flux Fill • How SAM 3 handles multi object detection with impressive accuracy • How SAM 3 can segment video, including a full moving vehicle example
This is one of the most capable segmentation models available today, and Griptape Nodes makes it easy to incorporate SAM 3 into your visual workflows.
This video demonstrates an example where a custom Nuke Gizmo is created to drive a Griptape Nodes Workflow, giving access to agentic generative AI from within Nuke.
Demonstrates the technique for designing the characters and uses a combination of models to get the results, including Flux 2, Nano Banana, Veo, Eleven Labs. Editing is accomplished with Adobe Premiere, and Griptape Nodes brings it all together.
This video demonstrates how to integrate OpenUSD into Griptape Nodes using an MCP server to interact with USD scenes through AI agents .
The process involves:
- Generating 3D Assets from Images: The video shows how to take a reference image, describe its contents, and then use Flex Image Generation to create 3D renders of each asset. These assets are then organized into a folder structure.
- Creating USDZ Assets with Rodin 3D: The generated image assets are converted into 3D USDZ assets using the Rodin 3D generator node, which is part of the Griptape library.
- Orienting and Scaling USD Assets: The video explains how initial USDZ models generated by Rodin often have incorrect orientations (Z-up) and are uniformly sized (1x1 cube). Griptape Nodes and an MCP tool are used to scale and orient these models correctly based on the reference image, saving them as clean USD assets with baked-in transform operations.
- Positioning Assets on the Floor Plane: The video demonstrates how to use the USD MCP tool to determine the bounding box of assets and then automatically adjust their Y-position so they sit flat on the floor. This can be done either through a nodal workflow or by chatting directly with the USD MCP server.
- Composing Scene Layouts: Griptape Nodes are used to build a workflow where an agent can interpret a reference image with a drawn axis to understand the relative positions and rotations of objects. This information is then used by the MCP task to create a USD layout file, positioning all the assets non-destructively.
- Creating Asset Variants: The video illustrates how to create a variant of an asset (e.g., an apple with a bite out of it). This involves generating the new 3D model and then using the MCP task to define a variant set within the original asset, allowing easy switching between different versions in applications like Maya.
Helpful Links: Autodesk Maya: autodesk.com/products/maya/overview OpenUSD: openusd.org/release/index.html Rodin3D: hyper3d.aiGetting started with Open Source Diffusion Models and Griptape NodesGriptape2025-11-03 | In this walkthrough, we’ll show you how to use open-source diffusion models inside Griptape Nodes — our flow-based editor for building and automating generative AI workflows.
🚀 Why Use Open Source Models? There are two major reasons creatives and developers choose to work with open models: 1 - Privacy & Control – Run inference on your own hardware and keep prompts, images, and IP completely private. 2- Diversity of Models – Access hundreds of community-trained image generation and enhancement models available on Hugging Face Diffusers.
🧠 What You’ll Learn - How to enable open-source image generation in Griptape Nodes - Using the Advanced Media Library and Diffusion Pipeline Builder - Downloading and managing models directly from Hugging Face - Setting up your Hugging Face API Token inside Griptape - Connecting your diffusion pipeline to a Generate Image Node - Running a complete local workflow with Stable Diffusion 3.5
By the end of this tutorial, you’ll know how to spin up a fully local image generation workflow — giving you full control, speed, and creative flexibility.
#Griptape #GriptapeNodes #OpenSourceAI #DiffusionModels #StableDiffusion #HuggingFace #AIArt #GenerativeAI #aiworkflowGetting Started with LoRAs for Diffusion Models in Griptape NodesGriptape2025-11-03 | In this tutorial, we’ll walk you through how to use LoRA (Low-Rank Adaptation) models with Flux diffusion pipelines in Griptape Nodes. LoRAs enable you to fine-tune large models quickly and affordably — perfect for customizing image styles, enhancing realism, or maintaining consistent character designs across generations.
🔍 What You’ll Learn: - What LoRAs are and why they’re game-changers for diffusion workflows - How to attach LoRAs using the new Flux LoRA File node - Where to find LoRA models on Hugging Face and how to connect .safetensors files - Trigger phrases & configuration for specific styles (e.g. “miniature world”) - Combining multiple LoRAs in a single pipeline for more control
Real examples using: 🧊 Anti-Blur LoRA: Reduces shallow depth of field in Flux 1-Dev 🎞️ Realism LoRA: Removes artifacts during upscaling for smoother, cleaner images Best practices for scaling resolutions, reusing pipeline nodes, and optimizing image quality
💡 Why it matters: With LoRAs, you can adapt models in minutes — no retraining required. That means faster iteration, unique creative styles, and professional-grade results directly inside Griptape Nodes.
#Griptape #GriptapeNodes #LoRA #AIArt #DiffusionModels #Flux #GenerativeAI #MachineLearning #AIWorkflowGriptape Nodes Template - Fill in the story!Griptape2025-11-01 | A video demonstrating a new template available for Griptape Nodes that uses Agents and ElevenLabs Audio nodes, Seedream Image Generation nodes, and your own imagination to create a custom Fill-in-the-blank Story - performed by a generated narrator!
:link: Helpful Links: How to Install Griptape Nodes: • How to Install Griptape Nodes in Under 2 M... Griptape Nodes: griptapenodes.com Griptape Nodes on Github: https://github.com/griptape-ai/gripta... Griptape Docs: https://docs.griptapenodes.com/en/sta... Griptape Discord: / discordSeedream and Seedance in Griptape Nodes!Griptape2025-09-12 | As promised, we have more Seedream and Seedance goodness for you.
Introducing Seedream image generation with support for the new Seedream 4.0 model and Seedance first frame to last frame video generation. These models are available in Griptape Nodes with your Griptape Nodes Credits.
In this quick demo, Ian walks you through how to: - Generate stunning photorealistic images from simple text prompts - Maintain character consistency across multiple images - Edit and customize details like clothing colors effortlessly - Create smooth, dynamic videos from still frames with Seedance
Imagine bringing your ideas to life with just a few clicks. Straight from concept to video!
Ready to try it yourself? Sign up now at griptapenodes.com and unlock these new capabilities today.
Watch the full demo here and see the magic in action!
AI #CreativeTech #ImageGeneration #VideoCreation #GriptapeNodes #InnovationTurn any video into a ready-to-post social media update!Griptape2025-09-11 | We're excited to share a quick demo of a simple yet powerful marketing automation flow using our latest feature enhancements in Griptape Nodes.
In this video, we walk you through how to: - Load video content - Extract audio from video with our new **Extract Audio Node** - Transcribe audio using the WhisperOne OpenAI model - Automatically generate engaging social media post copy from the transcription
This streamlined process saves you time and effort, turning your video content into ready-to-publish social media posts — all automated within Griptape Nodes!
Watch the full video to see it in action and learn how you can leverage this for your marketing workflows.
Ready to try it yourself? Sign up for Griptape Nodes today and start automating your content creation!
#MarketingAutomation #ContentMarketing #VideoMarketing #AI #GriptapeNodes #Automation #SaaS #OpenAIDirect access to ByteDance’s powerful Seedance video generation model!Griptape2025-09-09 | We’re thrilled to introduce **direct access to ByteDance’s powerful Seedance video generation model** — now available *right inside* Griptape Nodes! No more juggling extra libraries or managing separate API keys. Just seamless, integrated video generation at your fingertips.
What can you do with Seedance on Griptape Nodes? - Generate stunning videos directly from text prompts (text-to-video) - Create dynamic videos from static images (image-to-video) - Choose resolutions up to 1080p and multiple aspect ratios, including portrait aspect ratios that are perfect for social media (Instagram Reels, TikTok, Snapchat) - Generate 5 or 10-second clips with rich, high-fidelity detail
Ian walks you through real examples — from an ice cream truck melting with chocolate sauce to a horse galloping through a sunlit field, and even pyrotechnics lighting up a dance festival — showcasing how easy and powerful this new feature is.
Ready to try it yourself? Sign up for a Griptape Nodes account and get credits to start generating your own videos today!
#AI #VideoGeneration #GriptapeNodes #Seedance #ByteDance #MachineLearning #NoCode #TechInnovation #ContentCreation #SocialMediaMarketingFrom Image to Character: Creative Pipeline with Nano Banana Gemini and Griptape NodesGriptape2025-09-04 | 🔎 What you’ll see in this video: - Character views → Generate full-body, side, and back views from one image - 3D mesh generation → Export into Maya, Blender, or Unreal for refinement - Dynamic posing → Create multiple action poses (like singing on stage) from a single base design - Relighting & backgrounds → Instantly swap from indoor to outdoor stage settings with realistic lighting - Style & wardrobe changes → Experiment with different looks and clothing variations - Voice integration → Generate a character voice and soundtrack with Eleanor Knaps API - Upscaling & refinement → Polish final results with open-source upscalers, ControlNets, and Flux
💡 Why it matters: This flow is a complete character pipeline—going from concept art to production-ready assets in record time. Whether you’re building films, games, or virtual experiences, this method gives you speed without sacrificing creative control.
#Griptape #GriptapeNodes #GenerativeAI #3DModeling #CharacterDesign #AIAnimation #CreativeAI #VirtualProductionImage Generation with Griptape NodesGriptape2025-08-19 | Here's a tour of Griptape Nodes' image generation features, from simple setups to advanced workflows. See how it helps move from experiments to professional pipelines. Whether starting with AI image generation or exploring custom pipelines, this walkthrough is for everyone.
What you’ll learn: - Chat-based image generation → Use the built-in chat agent to create images and pull them directly onto your canvas. - Basic workflows → Generate images with models like DALL·E 3 and GPT Image 1, and see how prompt style impacts results. - Enhanced prompts → Automatically upgrade your prompts with LLM-powered prompt engineering for more stylized, professional results. - Custom prompt enhancement → Take control by defining your own enhancement rules and stylistic preferences. - Cross-model comparisons → Run the same prompt through multiple image generation models (OpenAI, Grok, Runway ML, etc.) for different creative outputs. - Reference images → Keep consistency across generations—ideal for product placement, advertising, or character design. - Looped workflows → Generate multiple variations across locations, styles, or parameters with iterative loops.
Tools & Models Featured: - OpenAI (DALL·E 3 & GPT Image 1) - Grok image model - Runway ML with reference image support - Black Forest Labs’ Flux models - Griptape Nodes’ prompt enhancement agents
#Griptape #GriptapeNodes #aiart #generativeai #promptengineering #creativeai #aiworkflowsGriptape Nodes - Snowflake Cortex IntegrationGriptape2025-08-15 | Due to popular demand, we've integrated Griptape Nodes and Snowflake Cortex.
With Griptape Nodes, our drag-and-drop generative AI workflow editor, and Snowflake Cortex, you can turn data into action—perfect for reporting, storytelling, or (in this case) visual insights.
💡 In this demo you will see: - How Griptape Nodes connects to Snowflake Cortex Search and Cortex Analyst as agent tools - How to query data directly from your Snowflake account - Crafting an Agent node that automatically generates image prompts from your data - Running the workflow locally in Python or deploying it to Griptape Cloud for API access - Integrating with a Streamlit app to create a live, interactive front-end for your hosted workflow
🛠️ Key technologies used: - Griptape Nodes – Flow-based editor for building generative AI workflows - Snowflake Cortex Search & Analyst – For querying and analyzing structured + unstructured data - GPT-4.1 Mini – LLM powering the workflow’s Agent - Streamlit – For building interactive front-end apps - Griptape Cloud – Deploying workflows as cloud-hosted APIs
📌 Pro Tip: Once deployed to Griptape Cloud, your workflow becomes an API endpoint you can call from any application—unlocking endless possibilities for automation and integration.
#griptape #GriptapeNodes #snowflake #cortex #generativeai #aiworkflows #datavisualization #aiautomation #llmHow to Install Node Libraries in Griptape NodesGriptape2025-08-15 | In this step-by-step tutorial, we’ll walk you through how to install and configure node libraries in Griptape Nodes—our powerful flow-based editor for creative professionals who want to integrate the latest generative AI models into their workflows.
In this video, you’ll learn: 📂 How to find and download third-party or open-source node libraries 🖥️ How to clone a GitHub repo and locate the library manifest file 🔑 Setting up API keys and secrets for libraries (like Black Forest Labs) ⚙️ Registering your new library in the Griptape Nodes engine 🎨 Testing your installation by generating stunning images with the Flux text-to-image node 💡 Pro tips for keeping your libraries organized and ready for creative projects
Why install libraries? Node libraries expand Griptape Nodes with new models, APIs, and creative tools—letting you go beyond the built-in nodes and customize your environment for your unique needs.
💬 Got questions or stuck on a step? Join our Griptape Discord and connect with our community of creators and developers.
#Griptape #GriptapeNodes #aiworkflow #generativeai #nocode #creativeai #BlackForestLabs #aigeneratedart #aiartProducing an Email Newsletter with Griptape NodesGriptape2025-08-11 | In this video, we show you how we use Griptape Nodes to build the actual Griptape Nodes newsletter.
📌 What you’ll learn in this demo: - How to create a reusable, parameterized text workflow in Griptape Nodes - Using past newsletters as style + format examples to train your agent - Pulling in new content sources like blog posts, product updates, or event pages - Attaching web scraper & search tools so your agent can find and read external copy - Building custom rule sets for tone, style, and formatting—then locking them for reuse - Merging structured inputs, examples, and instructions into one powerful agent prompt - Generating full newsletter drafts that are ready for polishing and sending
💡 Why this matters: If you create recurring content—newsletters, release notes, announcements—this workflow will save hours while keeping your brand voice consistent. Plus, it’s flexible: swap out URLs, rules, or templates and reuse the same flow for any audience or topic.
🚀 Try it yourself: Build your own text-based automation in Griptape Nodes and see how much time you can save.
🔗 Learn more at griptapenodes.comAnimating Images with Griptape NodesGriptape2025-08-07 | See how easy it is to take a static image and bring it to life using a reference video, with full control over body movement, expressions, and style!
🛠 In this demo, you’ll learn how to: - Use Griptape Nodes’ built-in chat to quickly refine your text prompts - Generate high-quality images with Flux.1 Krea (plus tips for quantization settings) - Import reference videos. Either from your own recordings or external assets. - Connect image and video inputs for realistic motion transfer (Runway Act-Two) - Adjust body control and expression intensity for perfect results (Runway Act-Two) - Play back your animated sequence, complete with synced audio - Add extra polish with upscaling, color grading, and other Griptape Nodes tools
💡 Why this is exciting: With this workflow, you can transform any still image into a fully animated performance, perfect for creative projects, short films, virtual production, and more. It’s quick, powerful, and only scratches the surface of what’s possible with Griptape Nodes.
🚀 Try it yourself when Flux tools drop, and let us know how you’re bringing your images to life!
Here's a brand-new productivity feature in Griptape Nodes: the ability to lock and unlock nodes.
This simple yet powerful tool enables you to skip processing on certain nodes, saving time, compute, and costs, while maintaining a fully connected workflow.
🎥 In this demo, you’ll learn how to: ✅ Lock a node to skip processing while still running the rest of your workflow ✅ Prevent edits to locked nodes (including properties & sidebar settings) ✅ Unlock nodes when you’re ready to run them again ✅ Speed up tests by avoiding reprocessing costly or time-intensive nodes ✅ Keep workflows connected—locked nodes can still send/receive data from other nodes
💡 Why it’s great: - Speeds up iteration and testing - Saves on compute and API usage - Protects important results from being accidentally overwritten
📌 Perfect for: - Complex workflows with expensive compute steps - Iterating on specific parts of your flow without rerunning everything - Keeping final outputs safe while experimenting downstream
We believe this will make Griptape Nodes even faster and easier to work with, particularly for power users handling large, multi-step projects.
Here's a full walkthrough of one of the most powerful visual tools we've added: the Image Bash Node.
Whether you’re a concept artist, filmmaker, or creative technologist, this tool gives you Photoshop-style compositing right inside Griptape Nodes—and it’s only the beginning.
🎥 In this demo, you’ll learn how to: ✅ Use the Image Bash Editor to layer, move, erase, and paint multiple visual elements ✅ Customize your canvas size, and fine-tune the layout for your scene ✅ Fix hard edges, mismatched lighting, and stylize assets using Flux + LoRA models ✅ Generate AI text prompts from your composition using the Describe Image node ✅ Pass your image through a ControlNet setup to stylize with precision ✅ Use Clear Reality upscaling to get beautiful 4K resolution ✅ Reapply models and blend styles for a cinematic final result ✅ Compare your original photo bash with the final AI-enhanced version
🌟 Example used: A custom composite from Amaru's short film Alas de Papel, reimagined and polished using trained LoRAs, multi-pass control, and AI-powered refinements.
💡 Why it’s powerful: This workflow lets you go from rough kitbash to final cinematic render using layers, open-source models, and intelligent automation—all within a single creative platform.
🖼️ It’s not just “generate and hope”—it’s intentional artistry + AI augmentation.
👀 Watch the full video to see how lighting, shadows, textures, and tone are completely transformed—while still keeping control in the artist’s hands.
Whether you’re working offline, need more control, or just love experimenting with custom models, this update lets you connect agents and image tools directly to LLMs running on your own machine.
🎥 In this demo, you’ll learn how to: ✅ Download and run models ✅ Use your model in a chat agent or an image description workflow ✅ Seamlessly connect nodes via PromptModelConfig in the Griptape canvas ✅ Run a local vision model to describe an image directly—no cloud required!
Set up everything from model download to workflow integration in under 3 minutes
💡 Why this matters: You now have full flexibility—run your favorite LLMs locally, integrate vision models, reduce reliance on cloud APIs, and maintain full control over your AI workflows.
Meet Amaru Zeas-Sigüenza — a multi-award-winning filmmaker, top 100 3D artist, and Creative Director at Griptape. With a background spanning film, VFX, games, and virtual production, Amaru has worked with giants like Microsoft, Amazon Game Studios, and AWS.
In his latest project, Alas de Papel (youtube.com/watch?v=QbjJSGQ8sH0) Amaru fused his deep creative experience with the power of Griptape Nodes and generative AI — and built an entire short film solo in just 14 days.
🛠️ In this video, Amaru shares: - How Alas de Papel was created using Griptape Nodes, Runway, Unreal Engine 5, Photoshop, and Substance Painter - How he used Griptape’s agent framework to generate 3,400+ images and 6,000+ render frames - How generative tools helped reduce a 300-hour process to just 11 hours—without sacrificing artistry - How he designed dynamic worlds with modular lighting, physics, and realism - Why multi-agent workflows, custom tooling, and AI collaboration are now essential to modern creativity
“You can tell when something is made in 45 seconds. It’s great for ideas, but not for the final pixel. I didn’t stop being an artist. I redefined what it means to be one with AI by my side.”
This isn’t just a workflow demo. It’s a creative manifesto for the next generation of artists, builders, and storytellers.
🚀 Griptape Nodes isn't just a tool. It’s your studio. Your canvas. Your co-creator.Creating Storyboards with Runway ML & Griptape NodesGriptape2025-07-30 | In this video, we’ll walk you through how to create a dynamic, image-based storyboard using Griptape Nodes and Runway ML’s powerful text-to-image models—perfect for creative teams in advertising, film, or content production.
🎯 Use Case: We're simulating the storyboard creation process for a fictional commercial promoting hiking holidays in the Scottish Highlands. This flow demonstrates how to build visually consistent scenes using a reference image and then expanding those scenes with prompt variations—all inside Griptape’s visual node editor.
🧠 What you'll learn in this video: ✅ Use Runway’s image generation models inside Griptape Nodes ✅ Generate a consistent visual style using reference images ✅ Apply text-to-image chaining to evolve and stylize characters or environments ✅ Create multiple scenes with forEach loops and prompt batching ✅ Organize your output using a custom-built Storyboard Node (grid layout, custom columns, padding, and size)
📣 If you’re in media, advertising, production, or just love advanced AI workflows—this one’s for you.
🔗 Helpful Links: Griptape Nodes: griptapenodes.com Griptape Nodes on Github: github.com/griptape-ai/griptape-nodes Griptape Docs: docs.griptapenodes.com/en/stable Griptape Discord: discord.gg/griptapeGriptape Nodes - WAN 2.2 Text-to-VideoGriptape2025-07-29 | We're excited to share that WAN 2.2 is now integrated into Griptape Nodes! This update unlocks powerful new capabilities for text-to-video generation using the latest from WAN 2.2’s A14B diffusers model.
🛠️ In this video, you'll learn how to: ✅ Install the WAN 2.2 model from Hugging Face ✅ Set up and run the WAN Text-to-Video node in Griptape ✅ Choose your model, add prompts & negative prompts, and generate stunning AI videos ✅ View results right inside the workflow ✅ Access WAN 2.2 in the Advanced Nodes Library
🎞️ Demo shown: A beautiful clip of foxes running through the woods—all generated using WAN 2.2 and a simple text prompt.
💡 Currently supported: ✅ WAN 2.2 Text-to-Video
🧪 Coming soon: 🖼️ WAN 2.2 Image-to-Video 🧠 WAN 2.2 Text-to-Video 5B
If you’ve already been using WAN 2.1 in Griptape Nodes, you’ll love the improvements in video fidelity and control with WAN 2.2.
🔗 Grab the model from Hugging Face, load it into your workflow, and let your imagination run wild.
🔗 Helpful Links: Griptape Nodes: griptapenodes.com Griptape Nodes on Github: github.com/griptape-ai/griptape-nodes Griptape Docs: docs.griptapenodes.com/en/stable Griptape Discord: discord.gg/griptapeGriptape Nodes - Dynamic Image DimensionsGriptape2025-07-17 | In this quick demo, we’re showcasing a powerful new feature added to the Display Image Node in Griptape Nodes—you can now extract an image’s width and height for use downstream in your workflows.
🔍 Why this matters: Many nodes—like Flux—require width and height as input parameters. Instead of typing them in manually or guessing, you can now automatically pull those dimensions straight from the image itself.
🎥 What you'll learn in this video: ✅ How to connect an image to the Display Image Node ✅ Where to find and enable the new width and height properties ✅ How to pass these values directly into other nodes (like Flux) ✅ How this improves dynamic, data-driven flows ✅ How to toggle dimensions on/off in the Properties panel
🧠 Bonus: If you change the image input, the dimensions automatically update—no need to reconfigure anything.
This feature helps streamline your workflow, reduces manual input, and unlocks more dynamic automation possibilities for image-based workflows.
Here's how you can combine the power of ChatGPT-4.1 image generation with contextual editing and open-source tooling—all inside Griptape Nodes.
💡 What’s in this video: ✅ Use ChatGPT-4.1’s Image directly in Griptape’s chat interface ✅ Instantly generate visuals with prompts ✅ Drag-and-drop images into your workflow canvas for seamless chaining ✅ Perform context-aware image edits ✅ Modify resolution, aspect ratio, and quality on the fly ✅ Apply open-source tools, LoRAs, ControlNet, and more
🛠️ Why this is powerful: - Chat is great, but context-based image editing gives you preservation + precision - You can now fluidly jump between prompt-driven creativity and structured visual workflows - It's modular, intuitive, and perfect for artists, builders, and AI tinkerers alike
📌 From here, the possibilities are endless. Build your own flows, style your images, integrate models, and bring your creative vision to life with zero-code flexibility!
We now have in-editor chat agent, which communicates directly with the newly integrated MCP server—no backend hacks, no glue code.
Yep, that's right. You may not be aware, but Griptape Nodes now features an integrated MCP server. This means that, aside from our sidebar agent, other AI applications that support MCP, such as Claude, can now interact with your Griptape Nodes with engine events, and do things like create an agent.
🎯 What you’ll learn in this video: ✅ How to define and register your own advanced library with schema version 0.2.0 ✅ How to use the new advanced_library_path to link a Python file with your logic ✅ Implementing before_library_nodes_loaded and after_library_nodes_loaded callbacks ✅ Real-time logging and execution when loading your library nodes ✅ How this sets the stage for powerful, modular, and library-scoped customization
💡 Why it matters: This is a powerful update for power users and dev teams. You can now extend Griptape functionality in your own libraries by adding logic, automation, or analytics hooks during library load time.
🔧 Example shown: - Loading a custom library with 99 nodes - Pre- and post-load events fire, printing key data - Now, devs can inject custom behavior into the Griptape loading process
This is just the beginning of a larger roadmap toward full library-level extensibility in Griptape.
🔗 Helpful Links: Griptape Nodes: griptapenodes.com Griptape Nodes on Github: github.com/griptape-ai/griptape-nodes Griptape Docs: docs.griptapenodes.com/en/stable Griptape Discord: discord.gg/griptapeExploring Custom Nodes for Stunning Image EffectsGriptape2025-07-02 | In this video, we're excited to showcase a creative deep dive into custom image-processing nodes built inside Griptape Nodes. These are perfect for adding depth, cinematic polish, and stylistic flair to your AI-generated visuals.
📦 What you'll see in this workflow demo: ✅ Depth Anything Mask Integration – Generate a depth map to give your image a sense of space ✅ Color Correction Node – Fine-tune brightness, gamma, and even tweak RGB tint values ✅ Image Blend Node – Inspired by Photoshop-style blending layers with adjustable opacity ✅ LUT Loader Node – Apply professional color grading using imported LUTs ✅ Lens Blur Node – Use the depth map to add realistic lens blur and simulate depth of field ✅ Film Look Generator Node – Add vignettes, grain, fade, exposure changes, and more for a cinematic finish ✅ Before/After Comparison – See the full transformation in one visual snapshot
🎯 Whether you’re building anime scenes, stylizing character renders, or just exploring the power of node-based image processing, this video highlights how Griptape Nodes can turn raw images into polished works of art, with a custom flow built for your vision.
🧪 All nodes featured are custom-built and extensible, meaning you can adapt them or build your own creative stack directly inside Griptape.
🔗 Helpful Links: Griptape Nodes: griptapenodes.com Griptape Nodes on Github: github.com/griptape-ai/griptape-nodes Griptape Docs: docs.griptapenodes.com/en/stable Griptape Discord: discord.gg/griptapeGriptape MCP Server Tool – Now Available in v1.8Griptape2025-07-02 | In this video, we'll be walking you through the new MCP Server Tool, introduced in Griptape version 1.8, which brings powerful interoperability between Griptape and external MCP (Modular Communication Protocol) servers.
💡 What is it? The MCP Server Tool enables seamless integration of any MCP server into your Griptape structures, serving as a standard tool that automatically translates MCP-native data types into Griptape artifacts, including text, images, and JSON.
🧪 In this demo, you'll see: ✅ Using the MCP Everything server to perform real operations (math, image retrieval, JSON reading) 🔄 Automatic conversion of responses into Griptape artifacts
🌍 Support for multiple transport layers: • Standard I/O • Streamable HTTP • WebSockets and Server-Sent Events as extensions
🎯 Why it matters: This tool unlocks endless integration possibilities with the extensive ecosystem of MCP servers available across the web, providing your AI workflows with real-time, modular communication power without reinventing the wheel.
🚀 What you can build: - Lightweight, plug-and-play integrations - Toolchains using external services via MCP - Dynamic workflows with multi-protocol communication
📌 Ready to explore? Start using the MCP tool today in your Griptape projects and expand what your agents can do.
In this video, we'll walk you through a real-world example using: ✅ Hosted Structures on Griptape Cloud ✅ Audio stem extraction using the Spleeter library ✅ Personalized Slack notifications via the Griptape Assistant ✅ And a fully orchestrated flow, built in Griptape Nodes
🎥 What you’ll see in this demo: - Load and process an audio file (song or rehearsal track) - Use a hosted structure to extract stems (vocals, drums, bass, etc.) - Iterate over each stem using a forEach loop - Auto-send a Slack message to each bandmate with their part (via your Griptape Assistant) - Preview a stem right in the UI
💡 Why this is awesome: You can now connect AI-powered backend services (like audio processing) with real-time communication and workflow automation—all from a visual, node-based environment.
⚙️ Griptape Cloud lets you: - Deploy & host your own Python code as "structures" - Invoke them via API from within any workflow - Use assistants to take actions (like messaging, summarizing, routing files)
📍 Check out our docs or GitHub to get started with the Griptape Cloud Nodes Library!
🧩 In this video, you'll learn about: ✅ A revamped onboarding experience with guided options for launching Griptape Nodes ✅ Upcoming cloud-hosted engines (choose standard or GPU-backed for your workload) ✅ Improved engine discovery and naming, so you can easily tell local from remote ✅ New session management that gives you control over engine communication ✅ Multi-tab syncing, allowing light real-time collaboration across browser windows ✅ A preview of how these updates lay the foundation for future collaboration features
🌐 Whether you're building locally or prepping for cloud deployment, these updates simplify your setup and unlock more powerful ways to build and manage your workflows, without the complexity.
Whether you’re generating images, calling APIs, or processing text, this update makes it way easier to handle lists and repeat actions without duplicating your flow.
🛠️ What you’ll learn in this video: ✅ How to use the forEach loop node to iterate over a list ✅ How currentItem and index work just like Python-style loops ✅ How to dynamically pass each item through a node ✅ How outputs are collected into a combined list using the forEach end node ✅ Why this dramatically reduces complexity and keeps your flows clean
🎯 Perfect for: - Looping over text prompts - Dynamic content generation - Batch processing - Simplifying repetitive flows
With this feature, you no longer need to duplicate flows to do the same thing multiple times. Just loop it—and keep building smarter!
🛠️ What’s new in the sidebar? ✅ Hierarchical organization – Related nodes are grouped logically (e.g. Agents, Prompts, Tools, Images, etc.) ✅ Library-specific views – Quickly explore nodes from individual libraries like Advanced Media ✅ Custom Favorites – Drag your most-used nodes into a personal Favorites section for fast access ✅ Resizable + Collapsible – Scale or minimize the sidebar to fit your workflow ✅ Spacebar Add Node menu upgrade – Favorites now appear right at the top for one-click access
🎯 Why it matters: As your projects grow, clarity and speed become essential. These updates bring powerful UX improvements that make node discovery and workflow-building easier, cleaner, and faster than ever.
🔗 Helpful Links: Griptape Nodes: griptapenodes.com Griptape Nodes on Github: github.com/griptape-ai/griptape-nodes Griptape Docs: docs.griptapenodes.com/en/stable Griptape Discord: discord.gg/griptapeExploring the Veo 3 Video Generation Model from Google AIGriptape2025-06-26 | We're diving into the Veo 3 model from Google AI—now accessible within Griptape Nodes. If you’ve been curious about generating AI-powered video with sound, this is the update you’ve been waiting for!
💡 What is Veo 3? Veo 3 is Google’s powerful video generation model capable of producing both video and audio simultaneously. While it’s currently behind a preview flag, you can request access through your Google Cloud account team and unlock it with proper credentials.
🛠️ What you’ll learn in this demo: - How to integrate Google’s Veo 3 model into Griptape Nodes - Setting up with your gcloud SDK or JSON credentials file - Using the Veo Video Generator node and Video Display node - Generating and rendering multiple videos from a single prompt - Viewing, downloading, and chaining outputs to other nodes (like upscalers)
🎬 In this video, we test it with a Star Wars-style prompt, generating two different Veo videos, each complete with motion and sound, right inside the Griptape Nodes environment.
📥 Get started: Head over to our GitHub repo and check out the Griptape Nodes library for Google AI. You’ll find full installation instructions, permission setup, and sample workflows: github.com/griptape-ai/griptape-nodes-library-googleai
🛠️ What’s new: We’ve optimized how workflows load, particularly when working with large and complex node libraries that have heavy dependencies. This change tackles one of the most common pain points we've heard from our community.
🎥 In this video, you’ll see: 📦 A side-by-side before and after of the loading performance 🔄 The same complex workflow loading noticeably faster ⏱️ Delays that used to be annoying… gone!
💡 Why it matters: As your AI projects scale, snappy performance becomes essential. This update ensures smoother development, quicker iteration, and a more enjoyable experience working in Griptape Nodes.
🌟 Why this matters: The integrated chat interface makes it super easy to build end-to-end AI workflows, from generation to annotation, without having to leave Griptape Nodes.
🛠️ What it does: 🧠 Converts natural language like “Christmas Eve 2027” into exact date formats 🔄 Supports a variety of preset and custom formats (ISO, casual language, military time, etc.) 🎛️ Easily switch between morning/afternoon or full custom time strings
📌 Examples from the demo: “Christmas Eve 2027” → 2027-12-24 “June 18th, 2024 at 3 p.m.” → June 18, 2024 in the afternoon Custom format → Friday, 0900 hours
🎯 Great for: - Logging and timestamping - Workflow scheduling - User-friendly date strings - Custom logic based on time of day
🔍 What this node does: ✅ Takes verbose agent responses 🔄 Feeds them into a new Summarize Text node 🧠 Returns a clean, concise summary so you get to the point—fast
📌 Example shown: An overly elaborate response from DeepSeek gets trimmed down to just what you need: “The agent chose 199,002 based on a calculated midpoint and adjustment.” Boom.
🎯 Perfect for: - Agent output cleanup - Making results more readable - Building cleaner UI responses - Summarizing long text for downstream tasks
If you’ve ever been unsure about whether your AI agent gave the right answer or if it aligned with your expectations, this tool is for you.
📊 What does this node do? It lets you score an agent’s response based on your input, expected output, and custom evaluation criteria. Whether you’re testing prompt quality, refining agent logic, or improving user experiences, this node gives you the feedback you need.
🔍 In this video, you'll learn how to: - Provide an input and expected output - Compare it to the agent’s actual output - Set custom evaluation criteria (e.g., factual accuracy, contextual fit) - Get a numerical score with a reasoned explanation - Use built-in presets for common evaluation types like paraphrasing, analogy, and factual correctness
💡 Example: Given the prompt “name something you might have at a kitchen table in the 1950s,” the agent replies “radio.” Is that a good answer? Evaluate it against “fork” with contextual criteria, and let the node decide.
📦 This is an awesome tool for: - QA-ing your AI flows - A/B testing outputs - Prompt refinement - Building smarter, more reliable agents
Today, we're diving into the newly released FLUX.1 Kontext model from Black Forest Labs—a powerful multimodal tool that lets you edit images with BOTH visual and text input, and even chain edits while maintaining visual consistency.
💡 In this video, you’ll learn how to: 🔧 Install the updated Black Forest Labs library in Griptape Nodes 🔑 Add your BlackForest Labs API key and register the model 🖼️ Use prompt + image together to edit visuals 🔁 Daisy chain multiple image edits (e.g., change car color → update background to cyberpunk city at night) 🎛️ Modify model selection and output dimensions on the fly
Whether you're a visual storyteller, AI creative, or just a curious builder, this is your chance to see how FLUX.1 Kontext and Griptape Nodes unlock a next-level workflow.
🆓 Griptape is free to use in both local and self-hosted environments. Learn more at griptapenodes.com
🔧 In this full demo, you'll see how you can: 📥 Drag & drop any image into your flow ✍️ Edit with natural language prompts like “add yellow sunglasses and a pink crown” 🔁 Use Flux LoRA for consistent and reliable edits 🔝 Upscale images with Clear Reality & Flux Post-Processing 🤖 Generate smart AI-powered text prompts with ChatGPT 🔍 Compare before & after images to see the transformation
This is an end-to-end image editing and enhancement workflow, powered by LLMs and LoRA models—no complex setup required. Whether you're a designer, AI tinkerer, or just love playing with creative tools, this video shows you how fast and easy it is to bring your vision to life.
How to get the model of the "Flux ICEdit" aka "DiptychFluxFillPipeline" node in griptape nodes: gist.github.com/dylanholmes/3a6c97343f1cc4d95ad69c1fdb4b4a38Griptape Nodes - New LoadGLTF NodeGriptape2025-05-27 | In this quick demo, we show off the LoadGLTF node in Griptape Nodes. Load, preview, and capture screenshots of your GLTF or GLB 3D models.
🎥 What this node can do: 🧩 Load any .gltf or .glb 3D model 🌀 Rotate and interact with it in real-time 📸 Save a snapshot of the current view with one click 🔍 View in fullscreen for better clarity 🔗 Use the snapshot downstream—like feeding it into the Describe Image node for AI-generated captions and analysis
Whether you're building creative tools, exploring AI-generated design workflows, or just want to see your favorite 3D models in action, this node is a game-changer.
Here's what it does: 🧠 Generates random text based on input—or creates it from scratch using AI 🔡 Choose from character, word, sentence, or paragraph modes 🔁 Use a seed value to control the randomness (great for reproducibility) ✍️ Pass in your own text OR let the node generate it using a built-in agent 💬 Dynamic content generation that adapts to your use case
Whether you're building an AI agent, testing a data pipeline, or just need to fill in the blanks, this node makes it stupid easy to get realistic, varied text with a single click.
🔥 Pro tip: When using sentence or paragraph mode with no input, it uses an AI agent to create something fresh based on the seed!
Here’s what it can do: 🔎 Input any block of text ✏️ Set a search pattern and replacement 🔁 See live, real-time output as you type ⚙️ Toggle case sensitivity, regex, and replace all 📄 Access built-in documentation with regex examples
Use it for quick text manipulation, data formatting, or just to see how many ways you can replace “capybara” with “watermelon” 🍉😄