How to Build AI Pipelines with WaveSpeed Desktop's Visual Workflow Editor

Most AI tools let you run one model at a time. Upload an image, pick a model, wait for results, download, re-upload to the next model, repeat.

WaveSpeed Desktop’s Workflow Editor changes that. It’s a visual, node-based canvas where you connect AI models, free tools, and media inputs into automated pipelines — then run everything in sequence or in parallel with a single click.

How It Works

The workflow editor uses a familiar node-and-wire interface (similar to ComfyUI, Unreal Blueprints, or Node-RED). Each step in your AI pipeline is a node on the canvas. Nodes connect through handles — the output of one node feeds into the input of the next.

Node Types

Node Type	What It Does
Media Upload	Load images, videos, or audio files as pipeline inputs
Text Input	Provide prompts, parameters, or text content
AI Task	Run any WaveSpeed cloud model (image gen, video gen, editing, etc.)
Free Tools	Use any of the 12 built-in tools (enhancer, converter, remover, etc.)
File Export	Save outputs to disk in specified formats
Preview	View intermediate results without saving
Annotations	Add notes and comments to document your workflow

Building a Workflow

Right-click the canvas or use the sidebar to add nodes
Drag nodes to position them
Connect output handles to input handles with wires
Configure each node’s parameters (model selection, prompts, settings)
Run the entire workflow or selected nodes

Example: Image Generation + Enhancement Pipeline

A common workflow that generates an image with FLUX, upscales it, and removes the background:

[Text Input: "A cat wearing sunglasses"]
    → [AI Task: FLUX Dev]
        → [Free Tool: Image Enhancer 4x]
            → [Free Tool: Background Remover]
                → [File Export: PNG]

Each node runs automatically when its inputs are ready. The entire pipeline executes with one click.

Example: Batch Video Processing

Process multiple videos through the same pipeline:

[Media Upload: video1.mp4, video2.mp4, video3.mp4]
    → [Free Tool: Video Enhancer 2x]
        → [Free Tool: Video Converter → MP4]
            → [File Export: /output/]

Execution Controls

Run Modes

Run All — Execute every node in the workflow from start to finish
Run Selected — Execute only the highlighted nodes
Continue From — Resume execution from any specific node (useful after fixing a failed step)
Retry Failures — Re-run only the nodes that failed

Batch Execution

Run an entire workflow multiple times (1–99x) with a single click. Each run can use different seeds or randomized parameters — ideal for generating variations or testing different prompts.

Real-Time Monitoring

The Execution Monitor panel shows:

Per-node status (queued, running, complete, failed)
Progress bars for long-running operations
Cost tracking per node and total workflow cost
Real-time output previews

Cost Management

The workflow editor includes built-in cost controls:

Real-time cost estimates — See the estimated cost before running
Per-execution limits — Set a maximum cost per workflow run
Daily budget tracking — Monitor your total spending
Cost breakdown per node — Understand which steps are most expensive

Canvas Features

Pan — Click and drag the canvas background
Zoom — Scroll wheel or pinch to zoom in/out
Fit View — Auto-zoom to show all nodes

Editing

Copy/Paste — Duplicate nodes or groups of nodes
Undo/Redo — Snapshot-based history (up to 50 states)
Context Menus — Right-click for quick actions
Multi-Select — Select and move groups of nodes together

Persistence

Auto-Save — Workflows save automatically as you edit
JSON Export/Import — Share workflows as .json files
Templates — Save common pipelines as reusable templates

Use Cases

Content Creation Pipeline

[Text Input: blog post prompt]
    → [AI Task: Image Gen (FLUX/Qwen Image)]
        → [Free Tool: Image Enhancer]
            → [File Export: blog-cover.png]
    → [AI Task: Video Gen (Wan 2.2)]
        → [Free Tool: Video Converter → MP4]
            → [File Export: social-clip.mp4]

Generate both a blog cover image and a social media video clip from the same text prompt.

Photo Editing Batch

[Media Upload: 50 product photos]
    → [Free Tool: Background Remover]
        → [Free Tool: Image Enhancer 2x]
            → [Free Tool: Image Converter → WebP]
                → [File Export: /processed/]

Process an entire product catalog: remove backgrounds, upscale, and convert to web-optimized format.

Video Post-Production

[Media Upload: raw-footage.mov]
    → [Free Tool: Media Trimmer (00:05-01:30)]
        → [Free Tool: Video Enhancer 2x]
            → [AI Task: Video Face Swap]
                → [File Export: final.mp4]

Trim, enhance, and apply AI effects to video footage in one automated pipeline.

Getting Started

Download WaveSpeed Desktop from GitHub Releases
Open the app and navigate to the Workflow tab
Add nodes by right-clicking the canvas
Connect them and configure parameters
Run your pipeline

The workflow editor is available on Windows, macOS, and Linux.

FAQ

Is the workflow editor free? The editor itself is free. Free tool nodes cost nothing to run. AI model nodes use your WaveSpeed account balance.

Can I share workflows with others? Yes — export any workflow as a .json file and share it. Others can import it into their WaveSpeed Desktop.

Is there a node limit? No hard limit on the number of nodes per workflow.

Does it work offline? Free tool nodes (converters, enhancers, background remover) work offline. AI model nodes require an internet connection to reach WaveSpeed’s cloud servers.

Is the workflow editor available on Android? Currently desktop-only (Windows, macOS, Linux). The mobile app focuses on the Playground and free tools.