SMTP quickstart
Get Sluice up and running in 5 minutes using SMTP.
Prefer a REST API? See the REST API quickstart for programmatic email sending with guardrails.
1. Sign up
Go to app.sluice.email and enter your email address. You'll receive a magic link — click it to create your account. Your organization is automatically set up and you're assigned the admin role.
No credit card required to get started.
2. Save your SMTP credentials
On first login, Sluice displays a one-time SMTP password. Save it somewhere secure — it can't be retrieved later (but you can regenerate it from Settings if needed).
Your credentials are:
| Setting | Value |
|---|---|
| Host | Your Sluice instance hostname |
| Port | 587 (STARTTLS) or 465 (implicit TLS) |
| Username | Your organization's email address |
| Password | The generated SMTP password |
3. Configure your outbound email
In Settings > Outbound SMTP, enter the credentials for the email provider you want Sluice to forward approved emails through. This is where emails ultimately get sent from.
Sluice supports:
- Gmail — SMTP with app password
- Microsoft Outlook — OAuth2 or SMTP
- Any SMTP server — SendGrid, AWS SES, Postmark, Mailgun, or any custom provider
Provider presets auto-fill the host and port for you. Hit Test Connection to verify everything works.
4. Point your AI agent at Sluice
Update your AI agent's SMTP configuration to use the Sluice credentials from step 2. This is the only change you need to make — no SDK, no API calls, no code changes beyond updating SMTP settings.
Here's what that looks like in a few popular setups:
Python (smtplib)
Node.js (Nodemailer)
n8n — In your Send Email node, set the SMTP host, port, and credentials to your Sluice values.
See the AI agent examples guide for more platforms.
5. Send a test email
Send an email through your agent. It will appear in the Review Queue on your Sluice dashboard — tuning mode is on by default, so all emails are held for review regardless of guardrail results.
Open the email in the review dashboard to see:
- The full email content (HTML and plain text)
- Guardrail results — which rules passed, which flagged concerns, and why
- Actions: Approve, Edit & Approve, or Reject
6. Tune your guardrails
Go to Settings > Guardrails to adjust your configuration:
- Enable or disable individual guardrails
- Adjust sensitivity thresholds (e.g., PII confidence scores)
- Write custom definitions for tone and content policy rules
- Upload a knowledge base for hallucination detection
Send a few more test emails and review the guardrail results. When you're confident in your setup, turn off Tuning Mode in Settings — safe emails will now auto-forward to your customers.
What's next?
- Learn about all 10 guardrails and how to configure them
- See AI agent examples for framework-specific integration guides
- Invite your team to review flagged emails