Hosted Email uses an in-house spam-scanning solution built on open-source technology. The system reduces inbound spam, phishing, and malware, and it limits outbound spam that could lead to IP blocklisting. This article explains how the filter scores messages, how per-user training shapes that scoring, and which headers you can read to interpret a verdict.
How the spam scanner makes decisions
Every message Hosted Email processes is given a spam score. The score is the sum of contributions from many rules and signals, including sender reputation, authentication results, content patterns, and per-user training data. A higher positive score means the message looks more like spam; a negative score means it looks more like legitimate mail (ham).
The filter also learns over time. When a user marks messages as spam or moves them between inbox and spam folders, the system adjusts how it scores future messages for that user. After enough training examples, per-user adjustments meaningfully influence verdicts on a per-mailbox basis.
Spam thresholds and scoring
Each user has a spam threshold that determines the score at which a message is treated as spam. Lower thresholds catch more messages but raise the chance of false positives.
- Normal (default): Messages scoring 9.00 or higher are treated as spam.
- High: Messages scoring 8.50 or higher are treated as spam.
- Very high: Messages scoring 8.00 or higher are treated as spam.
Note: The spam threshold setting must be enabled on the domain's brand before users see this option in webmail.
Mark an email as spam manually
Marking an email as spam reports the sender, headers, and content to the filter. Marking an email as not spam teaches the filter to treat similar messages as safe.
Tip: Reporting legitimate marketing emails as spam is generally ineffective — those senders are treated as trusted. Unsubscribe from their mailing list instead.
- In webmail, select the message and choose Spam.
- Press Accept to confirm the selection.
Note: The first time you mark a message as spam, you are asked whether to share data with the spam partner. If you accept, the prompt is suppressed for future actions. If you decline, the prompt appears every time.
Spam-related headers
The spam scanner adds several headers to each message. Reading these headers helps you understand why a message was scored the way it was.
- X-Spam-Status: Contains the overall spam score for the message.
- X-User: Shows the score adjustment applied by per-user training. Negative values push toward ham; positive values push toward spam.
- X-Rspamd-Server, X-Rspamd-Queue-Id, X-Stat-Signature: System headers used primarily by OpenSRS Support to diagnose delivery issues.
Next steps
- Tune the spam threshold for affected users. If end users report missed spam, increase their threshold from Normal to High or Very High. See Managing Missed Spam Messages.
- Configure allow and block lists. Override the filter for specific senders, domains, or IPs at the user, domain, or company level. See Set Allowed and Blocked Sender Lists.
- Diagnose rejections using the troubleshooting guide. If mail is bouncing or being refused at SMTP, work through the OpenSRS Hosted Email Spam and Rejections Guide.
Questions? Contact OpenSRS Support.
How helpful was this article?
Thanks for your feedback!
Do you still need help? If so please submit a request here.