Example chat

What it looks like

An example conversation. The store, products and results are made up.

Why a 3D studio is a good tool for agents

  1. 01

    Settings, not prompts.

    An agent changes a real parameter (warmer light, closer camera) and gets exactly that change. Nothing else moves.

  2. 02

    Same settings, same image.

    Results repeat, so an agent can compare variations fairly and learn from your ad data.

  3. 03

    Always the exact product.

    Your agent can make a hundred variations without one of them drifting off-brand.

mockivo · mcpExample
  1. You

    Set up my store. This is my URL: example-mugs.com

  2. Agent

    Connecting to Mockivo… Reading example-mugs.com.

  3. Found your products, logo, brand fonts and color palette.

  4. Agent

    Setup started. Two quick questions while it runs: should I start with the white mug and the magic mug? And which look fits your brand?

  5. You

    Yes, those two first. A warm, cozy kitchen. No neon.

  6. 3D models: in progress · Scenes: in progress

  7. Later
  8. Setup ready: 2 products in 3D, 3 scenes in your style.

  9. You

    Show me the white mug with our logo.

  10. Agent

    Here are 4 variations.

    • Variation 1
    • Variation 2
    • Variation 3
    • Variation 4
  11. You

    Number 3 is good. Make the light warmer and move a little closer.

  12. Agent

    Done. Same mug, same design: warmer light, closer camera. Nothing else moved.

    The edited variation: the same mug in warmer light, from a closer camera
  13. You

    Perfect. Make it a photoreal render, 4:5 for Instagram, plus an 8-second turntable.

  14. Agent

    Ordered. Both are in your deliverables queue.

  15. You

    What’s ready?

  16. Agent

    Deliverables: photoreal photo (4:5): ready · turntable video (8 s): rendering.

    Open deliverables →
  17. A week later
  18. You

    How did last week’s ads do?

  19. Agent

    I read your ad results. I tried 8 variations last week, and this one won: close-up, warm light, kitchen. I’ll make 8 more like it, with new angles and new copy.

  20. You

    Go.

  21. Agent

    8 new variations are ready for your review.

Capability map

One server between your agent and your studio

Your agent talks to one Mockivo MCP server. The server runs the studio tools for you.

01Your AI agentClaude, ChatGPT, Cursor, Antigravity or any MCP client
+Your other MCPsAd platforms and your site builder, connected in your agent
ReadyMockivo MCP serverSigns in to your Mockivo account and turns plain words into real studio settings.
03 · Studio tools
  1. read_store()Read your storeProducts, logo, fonts and colors from your URL.
  2. build_models()Build 3D modelsYour products in true 3D, ready for any scene.
  3. make_variations()Make variationsNew scenes and looks from your brand kit.
  4. edit_scene()Edit settingsLight, camera, lens, text. Change one, keep the rest.
  5. render()Render photos and videosPhotoreal stills and turntables in each channel’s format.
  6. deliver()Check and deliverFollow the deliverables queue and pick up finished files.
The studio tools your agent calls through the Mockivo MCP server.
  1. 01

    Connect and authorize

    Add the Mockivo MCP server to your agent and authorize your Mockivo account.

  2. 02

    Talk in plain words

    Your agent turns “warmer light, closer camera” into real studio settings. Change one, keep the rest.

  3. 03

    Close the loop

    With your ad platforms’ MCPs connected too, your agent reads what performed and makes more of what won.

Capabilities

What your agent can do

  • 01

    Connect and authorize with your Mockivo account.

  • 02

    Submit your store and check the setup status.

  • 03

    Answer setup questions about your products and style, so nothing is guessed.

  • 04

    Make variations from your brand’s knowledge base: your products, scenes, fonts and copy.

  • 05

    Order photoreal renders. Photos and videos, in the formats each channel needs.

  • 06

    Check the deliverables queue: setups, new models, new scenes and renders, with their status.

  • 07

    Add products and ask for new scenes, then make marketing material with them.

  • 08

    Connect ad results to the loop: add your ad platforms’ MCPs (Meta, Google, TikTok) to your agent. It reads what performed, then makes more of what won.

Who it’s for

  • Founders running several brands

    whose agent tests visuals for each one.

  • Marketers and agencies

    who already work through AI agents.

  • Developers

    building agents for e-commerce.

Frequently asked questions

Which AI agents does it work with?

Any assistant or agent framework that supports MCP, such as Claude, ChatGPT, Cursor or Antigravity.

Do I need the MCP server to use Mockivo?

No. Everything works in the Mockivo studio. The MCP server is another way in, for people who work through agents.

Can my agent read my ad results?

Your agent connects to your ad platforms through their own MCP servers, with your permission. Mockivo makes the creatives; your agent brings the results together.

Bringing your store to life.

Want your agent to run your studio?

Requesting access takes about two minutes. Building agents for e-commerce? Tell us in the last step.

Get access Log in

Private beta. Invites go out in waves.