Always on. Always yours.

Give it a computer.
It does the rest.

An AI agent that lives on your machine, talks to your team on WhatsApp/Telegram, and becomes an expert in your operations.

Request early access See how it works

Works on Mac, Windows, and Linux.

$ npm install -g leedab
$ leedab onboard
# That's it. Your agent is running.

Setup

Live in 10 minutes

No implementation. No engineering team. One person, one terminal, ten minutes.

1

Install

One command installs everything. No dependencies to chase, no Docker required.

npm install -g leedab
2

Connect

The setup wizard walks you through connecting WhatsApp, Telegram, Slack, or SMS. Pick your AI model.

leedab onboard
3

Chat

Your team messages the agent on WhatsApp, Telegram, or the web dashboard at localhost. It responds instantly and picks up where you left off.

Capabilities

More than a chatbot

LeedAB is an employee that never sleeps, never forgets, and never asks for a day off.

Computer use

Works like a human

LeedAB uses your computer like a team member would. Opens browsers, navigates systems, fills forms, reads screens, pulls reports. Accesses logged-in sessions the same way your team does. If a human can do it, LeedAB can do it.

Channels

Talks where your team talks

Your people don't learn a new app. They message the agent like a coworker.

WhatsApp
Slack
SMS
Telegram

Integrations

Connected to your tools

LeedAB works across your existing stack. No APIs to configure, nothing to rip out.

Local files
Gmail
Outlook
Calendar
Drive
Notion
CRM
ERP
Databases
Any website

Memory

Gets smarter with every interaction

Persistent memory across conversations. Never starts from scratch.

Local-first

Your data stays in your building

LeedAB runs on a machine you control. Your files, memory, and audit logs never leave your network.

Encrypted at Rest

All sensitive data stored on your disk with AES-256 encryption. Not in someone else's cloud.

Full Audit Trail

Every conversation logged locally. Your compliance, your machine, your rules.

Role-Based Access

Admin, operator, viewer. You control who can do what. Managed from the dashboard or CLI.

Management

Everything in one place

A web dashboard on your machine. No cloud login, no external dependency. Open localhost:3000 and you have full visibility and control.

  • Chat with your agent directly from the browser
  • Console showing agent status and channels at a glance
  • Sessions with full conversation history across every channel
  • Team management with admin, operator, and viewer roles
  • Settings for channels and access control

Questions

What is LeedAB?
An AI agent that runs on a machine you own. It connects to WhatsApp, Telegram, Slack, or SMS so your team can message it like a coworker. It has persistent memory, works across your existing tools, and handles routine operational work. Think of it as hiring someone who never sleeps and never forgets.
What hardware do I need?
Any modern computer with 8GB+ RAM. Mac, Windows, or Linux. A Mac Mini is ideal for always-on deployments because it's small, quiet, and energy-efficient. But a desktop, laptop, or Linux server works just as well. If you use a cloud AI provider, you don't need a GPU. If you want to run local models, a machine with a GPU or Apple Silicon is recommended.
Where does my data go?
Your files, memory, conversation logs, and audit trail stay on your machine. You can inspect every file. Nothing is stored on our servers. If you use a cloud AI provider, prompts are sent to them for processing. For maximum privacy, use a local LLM or a managed enterprise provider like AWS Bedrock or Google Vertex AI, which run within your cloud account.
Which AI models can I use?
Anthropic Claude, OpenAI GPT-4o, Google Gemini, DeepSeek, and OpenRouter (hundreds of models). You bring your own API key. For enterprise deployments, use AWS Bedrock or Google Vertex AI so prompts never leave your cloud account. You can also run local models for fully offline operation. Switch providers anytime with a single command.
How long does setup actually take?
About 10 minutes. One npm install, one setup wizard that walks you through everything (AI provider, messaging channel, team access), one start command. No engineering team required. No implementation project. No vendor calls.
Can I use this for my industry?
LeedAB ships with pre-built workflows for supply chain and logistics, but the agent is general-purpose. Any operations team that answers repetitive questions, processes documents, or coordinates across channels can use it. The agent learns your specific processes through memory.
How does memory work?
The agent maintains daily notes and long-term memory files on your machine. Tell it your team structure, your vendor names, your processes. It remembers across conversations and days. You control what persists. Memory is stored as plain files you can read, edit, or delete.
What happens if my machine turns off?
The agent stops responding until the machine is back on. Use pm2 or Docker to auto-restart on reboot. For true always-on, a dedicated Mac Mini or small server works great. We also offer cloud-hosted managed deployments if you'd rather not maintain hardware. Contact us and we'll set it up for you.
Do I need to be technical?
You need to be comfortable with a terminal (copy-paste commands). The setup wizard handles the rest. Once running, your team interacts through WhatsApp, Telegram, or the web dashboard. No code required to use it day-to-day.
I'm not comfortable giving an AI access to my computer. What are my options?
Completely understandable. A few options: dedicate a separate machine for the agent (a $300 Mac Mini works great) so it has its own environment with nothing else on it. You can also run LeedAB in a sandboxed or containerized environment using Docker, limiting what it can access. Or, contact us about our cloud-hosted managed deployment where we run and maintain everything for you on isolated infrastructure.

LeedAB

Always on. Always yours.

Your AI employee, running on a machine in your office. One install, ten minutes, and your team has an agent that remembers, works, and never clocks out.

Request early access