Free tool
What in this email looks like spam
Paste a subject line and body and this flags the wording patterns that make an email read as bulk: trigger phrases, too many links, image-heavy layouts, shouting capitals, stacked punctuation, and length. You get a risk verdict and specific fixes. It checks content only, never inbox placement.
Last reviewed 27 August 2026
Result
Content risk
High
Weighted score 19. Content only, not a deliverability prediction.
Length
53 words
50 to 125 words is our working band for a first touch
Checks
| Check | Found | Verdict |
|---|---|---|
| Trigger words in subject | None | Clear |
| Trigger words in body | free (money), guarantee (guarantee), no obligation (guarantee) | Fix |
| Links | 1 | Clear |
| Images | None | Clear |
| Exclamation marks | 2, 1 stacked | Fix |
| Shouting capitals | 1 word, 2% of the body | Fix |
| Length | 53 words | Clear |
| Subject | 35 characters, fake reply prefix | Fix |
Fix these
- Remove the Re: or Fwd: prefix. Faking a reply to someone who never wrote to you is the single most damaging thing in this email, and recipients report it rather than reply to it.
- Replace the trigger phrases in the body: free (money), guarantee (guarantee), no obligation (guarantee). Each one has a plain equivalent that says the same thing without the sales register.
- Cut to at most one exclamation mark, and never two in a row. Stacked punctuation is one of the oldest bulk-mail signals there is.
- Remove the words in full capitals. If a word needs emphasis, put it in a short sentence of its own.
- Add a plain opt-out line. US CAN-SPAM requires a working opt-out mechanism and a physical postal address on commercial email, and rules differ again in the EU and the UK. This is not legal advice, so check with someone qualified for your markets.
This reads wording only. Inbox placement is decided mostly by authentication, domain and IP reputation, list quality, and whether people reply, none of which this page can see. There is deliberately no percentage attached to the verdict, because any number here would be invented.
How this works
What the numbers mean.
- 01Words and phrases are matched against a categorised list of long-standing bulk-mail triggers, weighted by category. A hit in the subject counts double, because subject wording is examined more closely by both filters and humans.
- 02Structural checks run alongside the wording: link count against what the send type justifies, images against body words as a rough image-to-text proxy, exclamation marks and stacked punctuation, words in full capitals, and total length.
- 03Weights add into a single score with three bands. The bands are a rank ordering of risk, not a probability, and the tool says so rather than inventing a placement figure.
Assumptions and limits
- Trigger word lists are a blunt instrument. Modern filters weigh sender reputation and engagement far more heavily than vocabulary, so a clean score here is necessary and nowhere near sufficient.
- The image check uses the number you enter, because pasted text cannot reveal an image. The words per image figure is a proxy for the image-to-text ratio, not a measurement of it.
- No number in this tool predicts inbox placement, and we will not publish one. Any deliverability percentage you see attached to a wording checker is marketing.
Questions about this tool
Do spam trigger words still matter?
Less than they did, and more than zero. Filters now lean heavily on authentication, domain reputation, and recipient behaviour, so a well-authenticated domain sending wanted mail survives words that would have sunk it a decade ago. The words still matter to the human reading, which is the harder test anyway.
What actually decides whether my cold email lands?
Sender authentication with SPF, DKIM, and DMARC configured correctly, the reputation of the sending domain and address, how many recipients mark it as spam or delete it unread, and how many reply. Wording sits well below all of those, which is why this tool refuses to give you a placement percentage.
How many links should a cold email have?
One at most, and zero is often better. A first email whose only ask is a reply avoids the link question entirely and gets a higher response rate because replying is cheaper than clicking. Save the links for the second message, once there is a conversation.
Do I need an unsubscribe link on one to one cold email?
The rules vary by jurisdiction and by whether the message is commercial. US CAN-SPAM requires a working opt-out mechanism and a physical postal address on commercial email, and the EU and UK regimes are stricter about consent in the first place. This tool is not legal advice, so get the position confirmed for the markets you send into.
Why is a Re: prefix penalised so heavily?
Because it is a deception rather than a wording choice. It manufactures a relationship that does not exist, and recipients who notice it do not simply ignore the email, they report it. Reports damage sender reputation far more than any trigger word in the body.
This tool is free and there is nothing to sign up for. If you would rather have the work done than calculate it, that is what Zway does.