Trakto

Agent · inside the product

Ask for the work in plain text.
Get a document back, not a JPG.

Trakto's agent lives in the editor and in Flow. You describe the piece; it generates with your Brand DNA, creates variations, animates, and adapts formats — and it all opens in layers for you to finish. Consumes the plan's AI credits.

Create a 1080×1080 post for the summer line launch, with our brand applied, and give me 3 variations.
generate_design · brand kit: Solara · format: 1080×1080
create_variations · 3 directions
Done: 1 piece + 3 variations, with your Brand DNA's palette and typography. It's open in the editor — want me to animate the main one?

The conversation above is illustrative; the capabilities are what the product has today.

How it works

  1. 1

    Describe the piece

    A text brief — with format, goal, and, if you want, a reference image to recreate.

  2. 2

    The agent builds it on your brand

    Layout, text, and image generated with Brand DNA: palette, fonts, tone of voice. For carousels, 1 to 8 pages.

  3. 3

    You finish in the editor

    The piece arrives as a document with layers. Ask for variations, animate, adapt formats — or fine-tune by hand.

Who it's for

For anyone who produces pieces every day and wants to start from 70% done instead of a blank page — social, performance, brand teams, and agencies. The agent moves things forward; the editor guarantees the final result is yours.

What's actually real

What the agent does today.

Only what's in the product — each item consumes the plan's AI credits.

Generate design

From brief to editable piece, with your brand applied. Text stays text, image stays a layer.

Recreate a reference

Send a reference image and get the structure recreated as a document — to adapt, not copy.

Variations

Alternative directions for the same piece: composition, color, image, copy.

Animate

The static piece becomes motion with deterministic styles — no video credits consumed — and opens on the timeline.

Adapt formats

One design, every campaign size: feed, story, banner, DOOH.

Edit in conversation

"Swap the photo for a brighter one," "make the headline shorter" — in Flow, the copilot builds and adjusts the canvas.

Questions that reach us

FAQ — Agent

Is the agent on every plan?

From Starter up. On Free you use the editor's standalone AI tools within the daily limit; the agent comes with paid plans and consumes the monthly credits.

Is the result an image?

No — it's an editor document, with editable layers. Text, image, and shape stay separate so you can adjust them.

How many credits does it use?

Depends on the job: generating a piece costs like an image generation; animating with the native engine doesn't consume credits; AI video follows the chosen model's cost. The balance shows in the product before you confirm.

Is it the same as MCP?

The agent runs inside the product; MCP exposes similar capabilities to your own agent (Cursor, Claude). Both return an editable document.

Describe today's piece and watch the document open.

Start free No credit card · full editor