What is a custom agent here?

A saved sequence of tasks with your own instructions attached, run against a chosen persona, use case or competitor. It behaves like a repeatable workflow rather than a one-off prompt.

Do I need to write prompts?

No. You describe the outcome you want in plain language, pick the tasks it should chain, and Maestrix assembles the workflow. You can refine the wording afterwards.

Can agents run on a schedule?

Yes. Pro can schedule a workflow to run weekly or monthly per brand and deliver the output as a report.

One credit per task inside the workflow. Building and saving the agent itself is free.

Solution

Turn a workflow you repeat into an agent you run

Chain tasks, attach your own instructions, choose the inputs, and save it as something the whole team can trigger.

A custom agent is a sequence of real tasks, not a prompt macro. Each step reads the brain and hands its output to the next.

The same five steps, redone by hand every month

Research, persona refresh, angle set, campaign, report. Everybody knows the order and everybody rebuilds it manually.

Workflows that only exist in one person head
No way to run the same sequence across six brands
Quality that changes depending on who ran it

Tasks you can run here, and what they deliver

Every run reads your Marketing Brain first. One credit each, exports included.

Custom agentA saved multi-step workflow with your instructions~4 min · 1 credit

Workflow chainingOutput of each step passed into the nextper step · 1 credit

Step approvalRun all at once, or approve each stepper run · 1 credit

Scheduled runsWeekly or monthly execution with a delivered reportautomatic · 1 credit

Batch across brandsThe same agent run against every project you ownper brand · 1 credit

Browse the full catalog by discipline

Marketing Brain

What the brain contributes to every run

Stored objects go in, sharper objects come back out. Nothing starts from a blank prompt.

Reads from your brain

The project brain of whichever brand you run it against
Selected persona, use case or competitor as the agent input
Your own step instructions saved with the agent

Writes back after the run

Every step output stored against the project
Run history, so you can see what changed since last time
New objects discovered during the run

A look at the output

An excerpt from a real run, shortened for the page.

Custom agent · monthly competitive refresh1 credit · exportable

Input

Brand project, plus the two competitors already stored in the brain.

Step 1

Competitor scan on each rival. Pricing, messaging and page changes since the last run.

Step 2

Battlecard update, rewritten only where the scan found a real change.

Step 3

A short client-ready summary, delivered by email on the first Monday of each month.

The order it runs in

Strategy first, assets last. Each step feeds the next.

01
Describe
Say what the agent should achieve in plain language.
02
Chain
Pick the tasks and the order they run in.
03
Choose inputs
Persona, use case, competitor or brand.
04
Run or schedule
Trigger it now, or have it arrive monthly.

Why teams run it here

Repeatable, not remembered

The process leaves one person head and becomes a button.

Runs across brands

The same agent applies to every project you own.

Full control per step

Approve each step, or let the whole chain run.

Questions people ask

Run it against your own brand

Give Maestrix your site, get the first artifact back in minutes, and keep it in the brain.