Kreova
Turn any product into ads that convert in minutes.
Images, video, audio and AI actors from one API and one interface. Node-based workflows chain generations together, and a credit system meters every operation.
- Pricing
- From $29/mo
- Live at
- kreova.co
≈85
ads for $300
What a single $300 freelancer video buys in finished Kreova ads.
Minutes
turnaround
Typical time from brief to finished ad — freelancers quote 3–14 days.
25+
MCP tools
Agents like Claude Code or Cursor can run creations through the MCP server.
Why we built it
Every engagement ended the same way: the automation worked, then someone asked “can it also make our ads?” Agencies quote weeks and four figures per video. We built Kreova so a small team can test ten creative angles before lunch — and so we could learn what it takes to run a metered AI product in production.
Product intelligence — buyer, angles, and persona, mapped from a URL
What it does
- Product intelligenceDrop a product URL and Kreova maps the buyer persona, the angles worth taking, and hooks before anything is generated.
- Image adsOne brief produces three distinct compositions — you set the goal, platform, ratio, and tone.
- Video adsThree creative directions per brief, each with a hook, a preview, and the reasoning behind it. Pick one and render.
- Actor library200+ licensed AI actors — or train a custom one on your founder.
- Remix viral contentPaste a TikTok or Reel link and Kreova rebuilds its hook and structure around your product.
- Studio canvas + MCPChain text, image, voice, and video nodes into your own workflows — or let an agent drive 25+ tools through the MCP server.
How it works
Turn any product into ads that convert in minutes.
- 01
Create a project
2 min
Drop your product URL and visuals; the details are imported.
- 02
Learn the product
1 click
One click maps buyers, angles, and persona.
- 03
Set the outcome
1 min
Goal, platform, length, tone — with the exact credit cost shown before anything runs.
- 04
Render and post
Minutes
Finished ads in minutes. Failed or canceled runs refund automatically.
Kreova
Screens
Case study
How we built it
Playbook — messaging matched to a mapped persona
The problem
Generative AI products bleed money quietly. Every image, video and audio call has a different upstream cost, users expect instant feedback, and a single runaway workflow can burn a month of margin in an afternoon.
The approach
We built a credit ledger that meters every operation before it runs, with cost estimation exposed to the user up front. Workflows are a node graph, so a generation can feed the next step without a round trip through the client. Long-running jobs return an id and are polled, so nothing depends on an open connection.
The outcome
Predictable unit economics on a product where the underlying model costs change monthly, and a workflow engine that made new model integrations a configuration change rather than a release.
- 01
Prototype the pipeline
Weeks 1–4
A workflow canvas chaining text, image, voice, and video models — built for our own marketing first.
- 02
Meter everything
Weeks 5–8
Every node reports its cost; a credit ledger sits between the user and every run.
- 03
Make failure boring
Weeks 9–12
Automatic refunds, retries, and a render queue that survives provider outages.
- 04
Open the API
Ongoing
An MCP server with 25+ tools so agents can drive the studio directly.
Built with
- Next.js
- TypeScript
- Node.js
- PostgreSQL
- Modal
- ElevenLabs
Start with this offer.
Six questions, two minutes. We read every submission ourselves and reply within two business days.