Getting Started

This guide walks you through the Setup Wizard, your first login, and your first conversation with Gumm.

Prerequisite: Gumm must already be running. The fastest path is the one-command installer:
curl -fsSL https://raw.githubusercontent.com/gumm-ai/gumm/main/scripts/setup-server.sh | sudo bash
If you prefer manual Docker Compose, see the Installation guide.

1. Open the Setup Wizard

Navigate to http://your-server:3000 in a browser.

On a fresh installation Gumm detects that no configuration exists and automatically redirects you to /setup. You will see the six-step wizard.

2. Step-by-step wizard

Step 1 — Welcome

An overview of what the wizard will configure. Click Continue.

Step 2 — Brain Identity

Define who your assistant is. These values are injected into the system prompt on every conversation.

Field	Description	Example
Name	Your assistant’s name	`Gumm`, `Atlas`, `Jarvis`
Personality	A description of how it should behave	`A concise, technical assistant who prefers bullet points`
Rules (optional)	Hard constraints	`Always answer in English. Never make up data.`
Goals (optional)	High-level objectives	`Help me manage my projects and stay on top of my calendar.`

You can change all of these at any time from the Brain page (/brain).

Step 3 — Password

Confirm the admin password. This must match the value you set in GUMM_ADMIN_PASSWORD in your .env file.

Gumm is a single-user system. This password is the only credential for the dashboard.

Step 4 — OpenRouter API Key

Paste your OpenRouter API key. Gumm uses OpenRouter as a gateway to access LLMs (GPT-4, Claude, Mistral, Llama, etc.).

Create an account at openrouter.ai
Go to Keys and create a new key
Paste the key into the wizard

The key is stored encrypted in the database — it never appears in .env or logs.

Step 5 — Telegram (optional)

Connect a Telegram bot so you can chat with Gumm from your phone or desktop Telegram app.

To skip, just leave this step empty and click Continue.

To configure:

Open Telegram and talk to @BotFather
Send /newbot and follow the instructions
Copy the bot token (format: 123456789:ABC-DEF…)
Paste the token into the wizard
Optionally enter a Webhook URL (your public Gumm URL). If blank, Gumm uses long-polling mode, which works even on a private network

→ Full Telegram guide: features/telegram.md

Step 6 — Launch

Review your configuration and click Launch.

Gumm applies all settings and redirects you to the login page.

3. Log in

You’ll be redirected to /login. Enter the password you set in GUMM_ADMIN_PASSWORD and click Sign In.

4. Your first message

You’re now on the Chat page. The interface is a classic message input at the bottom — type your message and press Enter (or click Send).

Try something simple:

“What can you do?”

Gumm will describe its current capabilities, including any modules that are installed and enabled.

5. Explore the dashboard

Page	URL	Description
Chat	`/`	Your main conversation interface
Brain	`/brain`	Identity settings, memory, knowledge base, API keys
Modules	`/modules`	Install, enable, disable, and manage modules
History	`/history`	Browse all past conversations

Next steps

Chat guide — keyboard shortcuts, conversation management
Brain & Memory — how Gumm learns and remembers things about you
Modules — extend Gumm with new capabilities
CLI & Agent Mode — connect your machines and control them remotely
Telegram — chat on mobile via Telegram