open core · hosted cloud · Telegram-ready
EGGENT

An AI employee with
their own computer.

Not another chat window. Each one gets a machine of their own — files, a terminal, memory and a schedule — and does the work there while you get on with something else.

local-first · runs on ./data· Docker · Node · npm· managed cloud optional
WHO YOU HIRE

Hire the specialist the job needs.

You describe the job in plain language and the employee is set up for it. These are where people start — examples, not a menu.

▸ RAG

Knowledge specialist

Reads everything you have and answers out of it, with the sources. The documents and the vectors stay in your own database.

▸ MARKETING

Marketer

Keeps the content plan, writes the posts and publishes on schedule — every day, without being reminded.

▸ DESIGN

Designer

Turns a brief into a real page, then keeps sharpening it until it passes a review of its own.

▸ ANYONE ELSE

Whoever you need next

An analyst, a support desk, an assistant for one narrow process. Say what the job is and set them up yourself — no code.

BUILT-IN CAPABILITIES

A team of agents, one workspace.

Each project keeps its own chats, files, memory, schedules, skills and integrations — the desk one employee works at. Hire as many as the work needs.

▸ PROJECTS

Project-based workspaces

Group chats, files, skills, schedules, and memory per project. Context stays where it belongs.

▸ SKILLS

Skill-driven agents

Agents reach for the right skill pack for the job. Bundle company playbooks and workflows.

▸ MEMORY

Memory & knowledge

Agents remember across sessions and work with documents inside a project knowledge base.

▸ MCP

MCP server integration

Connect tools and data sources through MCP. Agents discover capabilities and use them when needed.

▸ SCHEDULES

Recurring automation

Schedule recurring tasks to any agent. Results land back in the chat thread.

▸ TELEGRAM

Telegram from anywhere

Create a cloud workspace from Telegram, then keep talking to the same Eggent bot on the go.

GET RUNNING

Self-host it, or deploy in 30 seconds.

~/ — one-command installer
$ curl -fsSL https://get.eggent.ai | bash# installs Docker if missing# keeps runtime state in ./data waiting for GET /api/health …✓ ok  http://localhost:3000$ 
EGGENT CLOUD

Managed, ready fast.

No Docker, no server, no model key required. Start with managed Eggent AI, then switch to your own provider whenever you want.

  • 14-day trial with hosted workspace
  • Managed Eggent AI included by default
  • Telegram onboarding and chat proxy
  • Upgrade path for teams and enterprise
HOW IT FEELS

Plain language in. Useful work out.

Eggent keeps the product surface simple while letting power users wire in files, schedules, MCP, memory, and custom skills behind the scenes.

01

Create a workspace

Use the browser or Telegram bot. Pick a workspace name, email login, and password.

02

Add context

Drop files into projects, add skill packs, and connect tools your agent can use.

03

Delegate

Ask for research, code, documents, checklists, reports, or recurring automation.

04

Keep going

Return from web or Telegram. Eggent preserves project chats, state, and memory.

READY

Build your first Eggent workspace.

Start hosted in seconds or self-host when you need full control.