Pixify feature

Workflow Editor

Visual DAG editor with 32 AI nodes — compose any pipeline

32 AI nodes across 5 categories — full modal coverage
Visual DAG, dependency-aware parallelism
Live progress via SSE streaming
Save as reusable templates — share publicly

Open Workflow Editor Explore all features

What is it

Pixify Workflow Editor is a React Flow-based visual workflow editor. Pipelines like "text → image → background remove → upscale → video → captions" that used to require 5 different tools now live on one canvas. Connect nodes with edges, hit Run, and the backend auto-parallelizes everything possible — 5-10× faster than doing it step by step manually.

How to use it

Get started in 6 steps

1
Start from a blank canvas
Click "New Workflow" for a fresh canvas. Left sidebar lists all 32 nodes, categorized: Input / AI Models / Processing / Output.
2
Start with an input node
Drag Image Input or Text Input to the canvas. Image Input uploads directly. Text Input takes a prompt.
3
Connect AI model nodes
Drag a Text to Image node, connect Text Input's output to its prompt port. Edges = data flow.
4
Chain processing nodes
Not enough? Add Upscale, Remove Background, Image to Video. Each node handles one step.
5
Hit Run
Top-right Run button. All parallelizable nodes execute concurrently (e.g., 3 t2i in parallel). Progress streams live to nodes via SSE.
6
Save as template
Like your workflow? Save As Template — to your library or publicly shared with all users.

Use cases

What other users build with it

Music video auto-gen

Lyrics → Split by Shot → batch t2i → i2v clips → Video Merge → BGM. Full MV in one click.

E-commerce pipeline

Product photo → Remove Background → multi-scene i2i → Upscale → File Output. One image, 20 scenes.

Consistent character series

Character Registry → Batch Render multi-prompt → Consistency Eval → pick the best.

Content localization

Video → Audio to Text → Translate → Text to Audio (multilingual) → Audio Video Merge. One pass, all languages.

Why Pixify

Real multi-step automation

Not a linear pipeline — a true DAG with branches, merges, parallelism.

32 nodes, all modalities

Image, video, audio, text, plus utility. Among the most complete workflow systems.

Live streaming progress

SSE pushes node state, progress, outputs to the canvas — no refresh needed.

Template ecosystem

Public template library + your private templates. One-click apply common scenarios.

Frequently asked questions

How is this different from the single-step tools?

Single-step tools (AI Image / Video / Avatar) are for one-off needs — open to result in < 1 minute, zero learning. Workflow Editor is for multi-step, reusable, automatable scenarios. Build once, reuse forever.

Can any node output connect to any input?

No — port types are strictly typed. Image ports only connect to image inputs, text to text. Prevents mid-execution failures from type mismatches.

Max nodes per workflow?

No hard cap, but past 50 nodes canvas performance degrades. Recommend splitting via Sub-workflow node references.

What happens on failure?

Failed nodes turn red with error message; succeeded node outputs are preserved. Fix and hit Continue to resume from the failed node — no full restart. Credits auto-refund.

Can I export workflows locally?

Yes. Export JSON saves nodes + edges + config to a file. Import JSON loads it onto any account's canvas.

Ready to start?

Open Workflow Editor