Wireflow is now a Claude connector. Run a batch from any Claude chat.
Set it up→Build the pipeline once. Call it from anywhere.
Wire up image, video, and LLM nodes on the canvas, then execute the workflow by reference from your code, Claude, or any agent. The graph is the function; the API runs it.
Free to start. No credit card required.
14,000+ creators and teams build on Wireflow.
Static 1:1
Carousel 1:1
Lifestyle 16:9Runs on Nano Banana Pro, Seedance, Veo, Kling, Seedream, and 95 more.
Start from an ad that already works
Pull the Meta Ad Library sorted by how long each ad has been running. Nothing runs for ninety days unless it is profitable. Pin one, take it apart, rebuild it around your product.
- Hook
- Problem said out loud inside two seconds
- Shots
- 6 shots, cuts at 0:02, 0:04, 0:07
- Beats
- Problem, proof, offer, ask
- Brief
- What to keep and what to swap
Every sampled frame drops into an image iterator, so one run redraws the whole shot list. Hand that to a video model and it is yours.
Searching costs no credits. Free accounts get 5 searches a day.
An endpoint out.
Design the pipeline visually, debug it node by node, then hit it with a POST. Same engine, same results, no glue code between model providers.
"CASA coffee, morning ritual hook, creator to camera, kitchen b-roll and pack shots."
UGC 9:16Creator 9:16
Static 1:1Carousel 1:1Lifestyle 1:1
B-roll 1:1Static ads get scrolled past
Your best static becomes a video ad without leaving the tab. Timeline editor, layered compositor, motion graphics, same canvas.
Generate every asset on one canvas
UGC, hook text, CTA, brand kit. Your team generates it all as nodes.
Pipe the layers into the compositor
Every generated layer lands as a branded frame on a transparent canvas.
Send it to the video editor
One click. The overlay lands on the timeline as a live layer, not an export.
Lay it over every UGC clip
One frame, a hundred creator videos. Swap the clip, ship the ad.
A real multi-track timeline in the browser. Trim, sequence, layer audio and captions, then export every ratio from one edit.
Stack product shots, text, and video on a layered canvas with GPU transitions and shader effects. What you see is the pixel-exact export.
Animated hooks, kinetic captions, logo stingers, and end cards, rendered to feed-ready MP4s in the cloud.
Your brand, wired into every workflow
Define your colors, logo, fonts, and product shots once. Every workflow starts on brand. Fork the kit and the whole team stays in sync.
ProductB-rollUGC 9:16Static 1:1
Carousel 1:1One generation. Every look.
Regenerating a video to fix its color burns credits every try. A regrade costs zero. Save the look to your brand kit and every output ships on brand.
Grade it once, then it posts itself on schedule.
Same generation. Zero re-inference.
Skip the media-pipeline plumbing.
26 nights. 25 finished carousels. Zero design hours.
A dating app our own team runs. One workflow picks the topic, writes the copy, and designs all six slides overnight. Nobody opened a design tool.
Read the case study
Explore what you can build.
Start from a production-ready template and run it in one click.
Browse all templates →
Images start around 4 credits and video around 10 credits a second. A typical mix lands nearer 30 credits an image and 17 a second. Credits roll over for 3 months on Professional and Team.
Questions from developers
Execute it by reference: POST to the workflow endpoint with an API key and your inputs, then poll the execution for results. The docs cover the request and response shapes.
Yes. Workflows are callable through MCP, so an agent can pick a workflow, pass inputs, and get media back without you writing per-model tool code.
Current image, video, and LLM models through one engine, including the FAL catalog. When a better model lands, you swap a node instead of rewriting an integration.
Workflow inputs are named ports. Your API call supplies values by name, the engine resolves them through the graph, and outputs come back on the execution.
Each node run is recorded with its status and output. You poll one execution and see exactly which step failed and why, instead of reconstructing it from logs.
Signing up is free with no credit card. Build workflows, browse templates, and earn credits through quick in-app actions to run your first generations. Plans start at $24 a month when you want volume.
The graph is the function.
The API runs it.
Build one pipeline on the canvas, call it from your stack, then hand it to your agent. Sign up free, upgrade when you want volume.