Intelligence,
integrated.

Connect your tools once. Let AI agents work across all of them.

See it connect →Acquire INT.bot

tool types

connection each

100%

actions logged

Illustrative figures

// connect

Watch an agent move across your tools.

Concept demo, scripted

INT0/5 linked

01 · CONNECT TOOLS

Click each tool on the map to link it to INT.

02 · PICK A SAMPLE TASK (connect all 5 first)

03 · AGENT LOG

$ waiting for task_

// api

One call. Every tool.

Illustrative API

from int_sdk import INT

int = INT(api_key="int_sk_...")

run = int.agents.run(
    "invoice-followup",
    tools=["payments", "crm", "mail"],
    require_approval=["mail.send"],
)
print(run.status, run.steps)

Illustrative API · not a real SDK

// why one layer

Built for agents that touch real systems.

One connection per tool

Authenticate each system once. Every agent reuses the same secure link.

Agents that share context

A customer in the CRM is the same customer in mail, docs and payments.

Every action logged and reviewable

Each read, write and send lands in an audit trail you can replay.

Humans approve sensitive steps

Payments, sends and deletes pause for a person to say yes.

// console

Agent console.

workspace / acme-sampleSample data

Agent	Status	Last run
Invoice Agent	running	2 min ago
Sales Recap Agent	idle	Fri 17:00
Records Agent	awaiting approval	14 min ago

// why the name

Three letters. Two meanings. One type.

INT▍

INT reads as intelligence and integration — exactly what the product does.

Developers already know int. It's a primitive: familiar, typed, fundamental.

Short enough for an API (int.agents.run), a command line ($ int run) or an app icon.

Intelligence,integrated.