Wireflow is now a Claude connector.

Set it up
Rizzly app iconCase study

Every night, one workflow ships tomorrow’s content

Rizzly is a dating app our own team runs. Its daily tips carousel is picked, written, and designed overnight by a single Wireflow workflow. By morning the finished slides and caption are waiting to post.

26
nights it ran unattended
25
finished carousels (one failed run)
6
designed slides per carousel
150
production-ready slides, zero design hours

26 nightly runs between June 29 and July 27, 2026, from production data. No reach or revenue claims: the posting side is a separate story.

The nightly loop

Brief in at midnight, finished carousel by morning

1

A nightly cron sends the brief

Rizzly’s server hits the Wireflow run API on a schedule. It picks a fresh topic, skips the last 14 themes so content never repeats, and passes the brief as a workflow input.

Run API with typed inputs

2

The workflow writes and designs

An LLM node turns the brief into slide copy with a hook, beats, and a payoff. The compositor lays out every slide in Rizzly’s look: photography, type scale, swipe cue. No template tool, no designer in the loop.

LLM node + compositor

3

Morning: finished and waiting

Rizzly polls the execution API, gets final PNGs straight off the CDN plus a ready caption, and queues the post. The same endpoints work for any app that can send an HTTP request.

Executions API + CDN output

Real output

These slides came off the workflow, not a design tool

Cover slides from three real nights, served from the workflow’s CDN output. The full carousel is six slides in the same look.

Carousel cover slide: She Stopped Initiating Texts
She Stopped Initiating Texts · July 27
Carousel cover slide: She Left You On Read
She Left You On Read · July 17
Carousel cover slide: Double-Text Without Desperation
Double-Text Without Desperation · July 15
She Left You On ReadDouble-Text Without DesperationSituationship Exit TextsDry Texter Revival MovesFirst Text After MatchingShe Keeps Rescheduling YouBreadcrumbing vs Interest TextsEnd the Talking Stage Fast

Themes the topic picker chose on real nights. It reads the last 14 before choosing, so the feed never repeats itself.

Want this running for your product?

We build the pipeline with you on a call, wired to your brand and your channels. Or open the workflow above and clone the pattern yourself.