A creative team that works your lanes. You approve.
Compose agents from skills, tools and a model. They report to managers, call each other, recreate proven winners from a hand-picked ad library — and every run waits on your yes.
Built from skills, tools and a model.
Every agent is yours to shape: a name, an @handle, a role, its own instructions and the model it runs on. Give it a manager and it joins the org chart — delegation and routing come with it.
- Attach skills (reusable playbooks) and pick from ~120 tools across studio, strategy, storyboard, video, ads, briefs and more
- Each agent runs on the workspace model by default — or its own
- Policy per agent: max turns, approval before running, a budget cap and max generations per run
- Add a Creative QA reviewer that checks the work, read-only
Name
Role
Handle
Used in Telegram to @mention this agent
An agent on every lane of the strategy.
Personas through to the Creative queue — ten lanes, each with an agent you can assign. Choose how much rope it gets, run it now or on a routine, and its output lands as drafts for review. Pick a lane to see its brief.
Template Scout · Creative ideas
Lane agent · daily routine
You approve each scheduled run before it starts; it may then generate up to 8 frames.
The 3am ceiling stare
DraftRecreated from #48213 · #50977
Proof before promise
DraftRecreated from #51102
One less thing on the nightstand
DraftRecreated from #47760 · #49318
They start from ads that already won.
Agents search the curated ad library for proven winners, recreate them for your market, niche and offer, and cite the ad ids they used — so you can see exactly what each draft was built from.
A hand-picked library
Proven ads and the formats that made them work — selected, not bulk-loaded. The Static Studio Crew’s Template Scout lives here.
Recreated for your offer
Winners become new concepts for your product and buyer, not copies — with the source ads cited on every draft.
Agents that call each other
A manager hands work to its crew; a reviewer checks it. Every step of every run is recorded, so you can see who did what.
Nothing ships without your yes.
Runs pause in the Inbox until you decide: approve a run, answer a question with a tap, accept a proposed data change, or green-light a paid generation with its estimated cost on the card.
- Approve, Reject, Reply — or Approve all
- Decide in the app, inline in a Loops chat, or on Telegram
- Telegram is live on your own bot: @loopy routes a task, @handle targets one agent
- Slack soon
Runs waiting on your approval before they continue.
Run Creative ideas lane
Approval · Template Scout · Daily routine · up to 8 frames
ReplyWhich angle leads the next test?
Input · Hook Writer · A question with tappable answers
Merge two overlapping desires
Patch · Ad Director · Proposed data change · 2 cards
ReplyGenerate 4 statics at 4:5
Spend · Template Scout · Paid generation · est. $0.32
Reply
On your own bot — connect it in Settings → Integrations.
Import a crew. Switch it on when you’re ready.
Template packs arrive with their org chart, skills, routines and goals already wired — and every routine switched off until you say so.
16 template packs
Agency Desk, Static Studio Crew, Performance Desk, Voice of Customer Lab, Video Ad Crew and more.
Routines
Daily, on an interval, manual, or on an event like “Idea approved” — and one routine can hand off to the next.
Every run on record
Each run is recorded and auditable — open its trace to see the steps it took, the tools it called and what it produced.
Your keys, your models
Bring OpenRouter, OpenAI, Anthropic or Google Gemini keys, and choose the model per agent.
Hire the crew once. Keep it forever.
Lifetime access for the first 40 — one payment, no subscription. Bring your own AI credits; we cover the app.