The take.
Filmed once. Eight seconds, one line, one product in hand. Everything below is this exact take with one thing changed.
Wireflow is now a Claude connector. Run a batch from any Claude chat.
Set it up→“okay this is the one I actually use every morning”
One take.
Every version of the ad.
Film the pitch once. Wireflow swaps the person, the look or the product on the same take.
Same take →
Loading the take…
Change one thing. Keep the take.
Same take. New everything.
Wireflow takes the video apart into the person, the outfit and the product, then puts it back together with new ones. The performance, the cut and the timing stay exactly as filmed.
New creator.
A different person says the same line on the same beat. Same room, same light, same hands.
New outfit.
The hoodie goes on mid-take and the pendant tucks under it. The performance does not move a frame.
New product.
Your jar, your label, in the hand that was holding theirs. Swap it once and every version updates.
New creator. His take.
The gym guy, filmed once with his shaker. The jar goes into the same hand on the same beat.
New creator. Her take.
Front seat, lip balm in hand. Swap it for the jar and the glance, the blink and the line all stay.
Start free, no cardEvery card is the hero's take, swapped, not re-shot.
What Wireflow is
Output
Ten ads. Keep the ones that would run.
Swipe right on the ones that would run, left on the ones that would not. That is the whole loop: keep, re-roll, ship. Tap the speaker to hear one.
Gym · shaker
1 / 10
Would this ad run? Swipe right to keep it, left to pass.
Make one with your productAbout $37 in credits for all ten. An agency quotes $150-$800 per variant. Free to start, no card.
Research node to ad
Start from
what already works.
Every version starts with an input. The Research node finds what is pulling on TikTok, the board reads the hook and the structure off it, and your creator makes the version.
Viral video
Analysisread off the source
Visual hook
Scroll-stopper0:02
Structure
- 0:00HookA question in the text, gummy to camera, no intro.
- 0:02Proof"About 30 mins later", eyes closing, one cut.
- 0:12PayoffOut cold. The product did the job.
The shape of a take that works, applied to your product and your creator. This card describes the ad on the right; the Research node hands the board the video, its caption and its stats to read the same way.
Your version
The last 10% is yours
Prompt it. Edit the last 10%. Say goodbye to CapCut.
The board makes the video from a prompt. The editor is for the last ten percent: nudge a caption, trim a beat, swap the sound. Change one word and every version updates.
See it in action
Nine-by-sixteen off the board: captioned ads and layered graphics, nothing touched by hand.
Rizzly · cover, compositor
Rizzly · cover, compositor
Rizzly · cover, compositorWork on it together
One board. Everyone on it.
Share the link and your client, your editor and Claude are on the same canvas at the same time. The ask, the fix and the approval happen next to the take, not in a thread of exports.
- Live cursors and edits. Nobody refreshes, nobody re-uploads.
- A view link for the client, an edit link for the team.
- Claude works the same board through the connector, so the version it makes lands on the node everyone is looking at.
Works inside Claude
Add Wireflow to Claude
Everything above, the takes, the swaps, the captions, the layered graphics, your agent can make from a message. The editor and the compositor are there for the last ten percent, when you want to move one thing by hand.
- 1
Copy the
connector URLYou will paste this URL into Claude in the next step.
https://www.wireflow.ai/api/mcp - 2
Go to Claude → Customize
In Claude desktop or claude.ai, go to Customize → Connectors and add a custom connector. Name it Wireflow, paste the URL, leave the OAuth fields empty. On Team and Enterprise an owner adds it under Organization settings first.
Open Claude Customize - 3
Connect and start creating
Sign in on the consent screen, then ask Claude:
Start creatingRun my captions workflow on https://cdn.wireflow.ai/landing/hero-pitch/base-540.mp4 and send me the output link.
The consent screen lists exactly what the connector asked for. Reading and running are granted by default. Tick Create and edit workflows on that screen if you want the agent to build and edit, not just run, because that permission is withheld unless you ask for it.
Leading AI models, one plan.
The best image, video and voice models on one board. No licences to chase, no tabs to juggle.
- Veo 3.1Google
- Nano Banana ProPhotoreal image generation
- Nano Banana 2Image generation and edits
- Veo 3.1Cinematic video with sound
- Gemini Omni Flash 1.1Image and reference to video
Seedream 5 ProByteDance Seed- Seedance 2.5The take, voice on
- Seedance 2.0First to last frame
- Seedream 5 ProStills and edits
GPT Image 2.5OpenAI- GPT Image 2.5Creator stills, pose-locked swaps
- GPT Image 2Detailed image editing
- Grok ImaginexAI
- Grok ImagineImage and video generation
- Kling O3Kling
- Kling O3Image to video, a start frame in
- Kling Video 2.6 ProImage to video
- Kling O1Video to video reference
Flux 2 ProBlack Forest Labs- FLUX 3Fast, sharp stills and image to video
- Flux 2 ProPhoto and creative images
- Hailuo 03 MaxMiniMax
- Hailuo 03 MaxReference to video
- WAN 2.7Alibaba
- WAN 2.7Reference to video
- ElevenLabs
- ElevenLabs TTSVoice
- Voice SwapYour voice, their take
- ScribeWords, timed
- Video Editor · captions and endcardWireflow
- Video EditorCaptions, endcards, the cut
- Topaz Upscale1080p and 4K
- Sync Lipsync v3Lips to a new line
- LTX Video 2.0Fast video
Recraft V4And more- Ideogram V4Type in images
- Recraft V4Vector and design
- Luma Dream MachineImage to video
One plan replaces the stack.
Seven of the many it replaces. Research, script, stills, video, voice, layout, the cut: one board, one bill, all of it wired together.
| Job | Wireflow |  Foreplay | ChatGPT | Nano Banana |  Higgsfield |  ElevenLabs | Canva |  CapCut |
|---|---|---|---|---|---|---|---|---|
| Research the ads that are working | ||||||||
| Write the hooks, scripts and copy | ||||||||
| Product stills, creator stills, swaps | ||||||||
| Takes and video swaps | ||||||||
| Voice and sound | ||||||||
| Layered graphics and carousels | ||||||||
| Captions, endcards, the cut | ||||||||
| Runs from Claude or your agent | ||||||||
| Start free | ||||||||
Swipe sideways to see every tool.
WHAT YOU SHIP
Finished creative, not a tool to learn.
Agency / freelanceWith Wireflow
New creative per week2-5Hundreds
Time to a new concept1-2 weeksMinutes
Cost per variant$150-$800Cents
UGC actorsCasting + usage feesBuilt-in roster
Iterate on a winnerRe-brief the agencyDuplicate & tweak
Channel-ready exportsManual resizingEvery ratio, one click
Getting it liveExport, download, post by handPublishes to your channels on schedule
“We ship commercial campaigns for global brands with zero shoot days. Wireflow is where the batch happens: variants overnight, and my team still hand-finishes the ones that matter. Nothing else kept up with our volume.”
Questions from people buying the ads
Yes. Upload product shots, logos, fonts, and a brand kit, and every variant is generated on-brand and stays consistent across the whole batch.
No. Meta does not restrict AI-generated ads, and does not require disclosure for commercial creative. What gets ads rejected is the same as always: claims, before and afters in restricted categories, and policy-violating text. Wireflow does not change any of that.
The roster is built for commercial use with no per-use talent fees, so you can run and scale UGC without licensing headaches.
1:1, 4:5, 9:16, and pre-roll for Meta, TikTok, YouTube, Google, and more, resized and named feed-ready in one pass.
Yes. Everything you create is yours to run, edit, and keep, on any channel, with no watermark.
Generate a full slate of hooks and formats, push to your channels, then duplicate and iterate on the winners without re-briefing anyone.
Yes. The Social Publish node sends finished media to your connected LinkedIn and YouTube accounts, right away or at a scheduled time. Posting is paced to protect the account: spacing between posts, a daily cap, and caption variation per platform.
Sign up, upload one product photo, pick a template. Most people have their first set of variants in under ten minutes and posting on a schedule the same day.
Signing up is free with no credit card. You can build workflows, browse every template, and earn credits through quick in-app actions to run your first generations. Plans start at $24 a month when you want volume.
Just want the finished ads? We run the board for you: your best ad rebuilt ten ways a week, creators, hooks, every ratio. From $5,000 a month. Bring one product.
See the offerBring one
product photo.
Pick a format, hit run. Free to start, starter credits included, upgrade when you want volume.
- Free to sign up
- Earn starter credits
- No credit card
- Your assets are yours to use
- Cancel anytime