--- title: "Your First Hour With Your Own n8n: A Setup Checklist" description: "A practical first-hour checklist for your new n8n instance: rotate password, enable 2FA, verify webhooks, pick an AI model, and run your first workflow in under 60 minutes." canonical: https://agentroost.app/en/blog/first-hour-with-your-own-n8n-checklist date: 2026-05-13T20:00:00Z --- [Canonical URL](https://agentroost.app/en/blog/first-hour-with-your-own-n8n-checklist) # Your First Hour With Your Own n8n: A Setup Checklist Your n8n editor is open at `https://.agentroost.app`. You have a blank canvas, a blinking cursor in the search bar, and exactly zero workflows. That's the right moment to spend 45–60 minutes on the setup tasks that never feel urgent until they are — rotating credentials, turning on 2FA, verifying your public webhook URL actually reaches the internet, and firing the AI node once so you know the included credits are live. This checklist runs through each step in the order that makes sense. Skip nothing in the first four; the rest are best-practice but worth doing before you build anything real. --- ## Step 1 — Change Your Password Right Now Your instance shipped with a bootstrap password. It's in your AgentRoost dashboard and it was transmitted over HTTPS, but it's still a one-time credential you should replace immediately. Inside your n8n editor: **Settings → Users → (your account) → Change Password**. Pick something you haven't used elsewhere. n8n stores passwords hashed, but the point is that your instance lives at a public subdomain — there's no IP allowlist between your editor and the internet. A strong unique password is your first line of defence. --- ## Step 2 — Enable Built-in TOTP Two-Factor Authentication n8n has native TOTP support. This is not a plugin or a paid feature — it's built in and works out of the box. **Settings → Personal → Two-factor authentication → Set up 2FA** You'll get a QR code to scan with Google Authenticator, Aegis, 1Password, or any TOTP app. Save your backup codes somewhere offline. Once enabled, every login prompts for the 6-digit code in addition to your password. This one step closes the most realistic attack vector against a public-URL workflow editor. --- ## Step 3 — Confirm Your Public Webhook URL Before you build anything that receives data from the outside world, verify the public webhook base URL is what n8n thinks it is. **Settings → General → Webhook URL** You should see `https://.agentroost.app/`. If it's blank or shows `http://localhost`, webhooks will generate broken URLs and external services (Stripe, GitHub, Slack, etc.) will get 404s. On AgentRoost this is pre-configured correctly, but it takes five seconds to confirm. Copy that URL — you'll paste it into every external integration you set up. **Quick smoke-test:** Drop a **Webhook** node on the canvas. Set it to `GET`, copy the test URL it generates, paste it into a new browser tab, and hit Enter. Switch back to n8n — you should see "Waiting for test event…" turn into a green execution with the empty GET payload. Done. --- ## Step 4 — Pick Your Default AI / Chat Model This is where AgentRoost differs from every other n8n hosting option: **you have AI credits already included**. You don't need an OpenAI API key, an Anthropic account, or any external billing setup. The AI nodes are wired to your included credits from the moment your instance starts. To set a default model: 1. Open **Credentials** (top-right menu → Credentials). 2. You'll find a pre-created credential called something like **AgentRoost LLM** (an OpenAI-compatible endpoint pointing to the AgentRoost LLM proxy). 3. That credential gives you access to 350+ models — GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro, Llama 3, Mistral, and more. Switch the model string in the node itself whenever you want. In any **AI Agent** or **Chat Model** node, select that credential and type the model ID you want in the Model field. A reasonable default for most workflows: ``` openai/gpt-4o-mini ``` It's fast, cheap on credits, and handles most classification, summarisation, and light reasoning tasks well. Swap to a larger model (e.g. `anthropic/claude-3-5-sonnet`) only when the task needs it — you'll burn credits more slowly this way. --- ## Step 5 — Run Your First AI Node (The Sanity Check) Don't build a real workflow before you've confirmed the AI node actually fires. This takes two minutes: 1. Create a new workflow. 2. Add a **Manual Trigger** node (click the `+` button, search "manual trigger"). 3. Connect an **AI Agent** node to it. In the Agent node, attach your AgentRoost LLM credential and set the System Prompt to something simple: `"You are a helpful assistant."` 4. Add a **Chat Trigger** input or a **Set** node that feeds the agent a fixed prompt: `"Say hello in three words."` 5. Click **Test Workflow**. You should get a response from the model within a few seconds. If you do, your credits are live and the AI pipeline is working end to end. If you get a 401 or a timeout, open a support chat — something unusual happened during provisioning. --- ## Step 6 — Set Up Your Timezone and Execution Log Retention Two settings that are easy to forget and annoying to debug later: **Settings → General → Timezone** — Set this to your local timezone. Every **Schedule Trigger** node (`0 9 * * 1-5` = weekdays at 9am) interprets cron expressions in this timezone. If it's wrong, your "9am Monday report" fires at 2am. **Settings → General → Execution History** — By default n8n keeps execution logs. Set a retention window that fits your usage (7–30 days is typical). Large workflows with many executions can accumulate data quickly; don't discover this months later. --- ## Step 7 — Create Your Folder Structure If you're new to n8n, this feels premature. Do it anyway — you'll thank yourself after workflow number fifteen. On the **Workflows** home screen, click **Add Folder** (top-right). A structure that works for most people: ``` 📁 Production 📁 Staging / Testing 📁 Templates 📁 Archive ``` Move every workflow to `Staging / Testing` while you're building it. Only move it to `Production` when you've run it at least once end-to-end with real data. "Archive" is for old versions you're not ready to delete. --- ## Step 8 — Invite a Second User (Optional but Recommended) If you're running this for a team or want a backup admin account: **Settings → Users → Invite User** You can assign the `Admin` role (full access) or `Member` (can build and run workflows but can't change instance settings). Credentials are still shared by role, so admins see all credentials, members see only what they created. Having a second admin email on file means you're not locked out if you lose your TOTP device before you've saved backup codes. --- ## How to Get Here on AgentRoost If you haven't started yet, here's the actual path: 1. Sign up at [agentroost.app](https://agentroost.app) (email/password, Google, Microsoft, or Discord). 2. Go to **Workspaces → New Workspace**, choose the **n8n** framework. 3. Name your instance. Your private editor opens at `https://.agentroost.app` within about two minutes — SSL already configured, AI credits already included, no Docker, no CLI, no API key setup. 4. Walk through the eight steps above. Plans start at $19.99/mo, which bundles the server and your AI/LLM credits into a single line item. There's a 14-day money-back guarantee if it doesn't click for you. [Compare plans](/en/pricing) · [See the n8n workspace](/en/agents/n8n) --- ## What You've Accomplished By the end of this first hour your instance is: - **Secure**: unique password + TOTP 2FA, no bootstrap credentials in use. - **Verified**: webhooks resolve to a real public HTTPS URL. - **AI-ready**: a working LLM credential with included credits, confirmed with a live test. - **Organised**: timezone set, folder structure in place, execution retention configured. That's the foundation. Everything you build from here — scheduled reports, Slack bots, data pipelines, AI agents — runs on top of it. Start with the workflow you've been putting off.