Do I need to know how to code to build an agent?

No. You create an agent from a button, choose a role from a list and write its job in plain English. Developers can also run agents on their own machine through the desktop app, but nothing on this page needs it.

How is this different from Zapier, Make or n8n?

Those tools run a fixed sequence you designed. A Burrak agent decides the sequence itself from your goal, which suits messy work like research, writing and triage, and suits less well work that must be identical every time. We wrote up the trade-off in “AI workflow automation: agent or workflow builder?” on the blog.

How many agents can I build?

Local agents on your own computer are unlimited. Remote agents, which run in the cloud 24/7, are capped by plan: 2 on Free and Pro, 5 on Pro Plus and 10 on Max.

The account is free on iOS and Android with pay-as-you-go credits. Paid plans start at $9.99 a month with 12,000 credits included, and the first week of any paid plan is $1. Credits pay for the AI model and for cloud runtime, and an idle agent costs nothing.

AI agent builder

A no-code AI agent builder: describe the job, not the flowchart

Create an AI agent in a couple of minutes. You write what you want done; the agent works out the steps.

Get started See pricing

Most “AI agent builders” are workflow editors with an AI step in the middle: you still drag boxes, wire triggers and decide every branch in advance. Burrak works the other way round. You create an agent, tell it what kind of work it does, connect the tools it needs, and then brief it the way you would brief a new hire. It plans the steps, uses the tools, checks its own result and reports back.

Each agent has its own name, chat history, memory and role, so you can keep several side by side, such as a researcher, an inbox assistant and a weekly-report writer. Remote agents run in the cloud around the clock, so a job you hand over at 6 pm can be finished when you wake up.

Things you could ask for

Written the way you would actually say them. The agent plans the steps itself.

“Every Monday at 8am, research what our three main competitors announced last week and send me a one-page summary.”
“Go through today’s unread email and tell me which threads need a reply from me, with one line on each.”
“Take the notes from this call and turn them into a proposal in the format of the last one I sent.”
“Check our Stripe account every Friday and tell me revenue, new customers and anything that looks unusual.”
“Find ten local businesses in our area that might need our service, with a contact and one reason each.”

What you get back

Finished work, not a chat transcript: a document or spreadsheet you can download, drafts waiting in your mailbox, an updated record in the tool you connected, or a short report. Every run keeps a full transcript of what the agent did, so you can see how it got there.

Tools it works in

Connect a tool once by signing in to it, and the agent works on your real email, calendar and files instead of whatever you paste into a chat.

GmailAn agent that reads your inbox, drafts the replies and sends what you approve.Google CalendarAn agent that reads your calendar and acts on what it finds there.Google SheetsLet an agent read, write and reason over your spreadsheets.Tavily Web SearchGive an agent the live web, with the links to check it by.StripeAsk questions about your revenue and get an answer, not a dashboard.ZapierReach thousands of apps that have no MCP server of their own.

How to set it up

1Create a free account in the iOS or Android app, or pick a paid plan for the web and desktop apps (your first week is $1).
2On the Overview page, create an agent and give it a name. Choose a remote agent if you want it working while your computer is off.
3Optionally give it a role from the Marketplace, such as Chief of Staff, Content Creator or Financial Analyst, which shapes how it approaches every task.
4Connect the tools it should work in, like Gmail, Google Calendar, Sheets or Notion, from the Connectors tab.
5Brief it in chat, or put the job on a schedule from the Tasks page (“every weekday at 9am”). You are only charged while it is actually working.

Roles that suit this work: AI Chief of Staff, AI Content Creator, AI Financial Analyst.

Where it is the wrong tool

It is not a visual flow builder. If you need the exact same steps in the exact same order every time, with every branch drawn out for an auditor, a workflow tool such as Zapier or n8n is the better fit, and an agent can call Zapier through its connector when you want both.
It does not build customer-facing chatbots to embed on your website, and it does not answer phone calls. The agents you build work for you and your team.
An agent reads instructions generously. Anything outward-facing, like sending email to customers, deserves a “draft it first” in the brief until you trust the result.

Common questions

Do I need to know how to code to build an agent?: No. You create an agent from a button, choose a role from a list and write its job in plain English. Developers can also run agents on their own machine through the desktop app, but nothing on this page needs it.
How is this different from Zapier, Make or n8n?: Those tools run a fixed sequence you designed. A Burrak agent decides the sequence itself from your goal, which suits messy work like research, writing and triage, and suits less well work that must be identical every time. We wrote up the trade-off in “AI workflow automation: agent or workflow builder?” on the blog.
How many agents can I build?: Local agents on your own computer are unlimited. Remote agents, which run in the cloud 24/7, are capped by plan: 2 on Free and Pro, 5 on Pro Plus and 10 on Max.
What does it cost?: The account is free on iOS and Android with pay-as-you-go credits. Paid plans start at $9.99 a month with 12,000 credits included, and the first week of any paid plan is $1. Credits pay for the AI model and for cloud runtime, and an idle agent costs nothing.