SendPulse Review 2026: A Third of Its Mail Went to Spam, and It Publishes No Number of Its Own
The Best: Six channels on one account, published API limits and an official open-source MCP server
The Worst: 62.8 percent inbox placement, fourteenth of the fifteen platforms tested
The short version
SendPulse does more things than almost any platform we cover. Email, an SMTP relay, chatbots across six messaging networks, SMS, web push, a CRM, landing pages, pop-ups and an online course builder, with a free tier that includes most of it and an official open-source MCP server on top.
And in the only independent test it has ever been included in, 62.8 percent of its mail reached the inbox. That was the January 2024 seedlist round, the twelfth and last before the panel was discontinued. Fifteen platforms were tested. SendPulse came fourteenth. The published verbatim: “SendPulse is the newcomer in our email deliverability tests and while we can’t provide data for their average of the last 3 rounds, they came second to last with 62.8%.”
Thirty-three percent went to spam. For comparison, ActiveCampaign in the same round reached 94.2 percent and MailerLite 89.8.
The part that decided this review is not the number itself. It is that SendPulse publishes no delivery-rate figure of its own. We looked: the glossary page on delivery rate carries no vendor claim, the features pages carry none, and nothing on the site states a percentage. Almost every competitor at this size advertises one. SendPulse is silent, and after reading the independent figure that silence looks less like modesty and more like a decision.
The 62.8 percent, and what it does and does not say
This is the most consequential figure in the review, so here is exactly what stands behind it.
| Round | January 2024, the twelfth round, and the last one ever run. The panel has since abandoned seedlist testing in favour of feature assessments, so no newer measurement exists or will. |
|---|---|
| Method | Roughly 80 seed addresses across 28 mailbox providers, with SPF, DKIM and DMARC configured. Definition quoted: “A correctly delivered email is defined as an email that arrives in the recipient’s inbox, which includes the promotions folders”. Promotions counts as inbox, which makes the score generous rather than harsh. |
| SendPulse | 62.8 percent inbox. 33.0 percent spam. 4.2 percent never arrived. |
| Rank | Fourteenth of fifteen. Only Benchmark, at 47.1 percent, was worse. |
| Context | ActiveCampaign 94.2 · Constant Contact 91.7 · GetResponse 90.9 · Moosend 90.1 · MailerLite 89.8 · Mailchimp 89.5 · Brevo 88.3 · Drip 88.2 · Mailjet 88.0 · AWeber 83.1 · HubSpot 77.7 · Omnisend 75.1 · SendPulse 62.8 · Benchmark 47.1 |
The panel publishes a per-round breakdown alongside the headline score. For SendPulse, that breakdown does not add up: extracted three times identically, its four cells sum to 106.8 percent, where every other row in the same table sums to about 100. The headline figure is computed as main inbox plus tabs, and we verified that formula holds for other rows. For SendPulse, 62.8 plus 33.0 spam plus 4.2 missing equals exactly 100.
So the headline 62.8 percent is sound and the spam and missing figures are sound. One cell in the breakdown is corrupt and we do not publish the split. We are telling you this rather than quietly reproducing a table we know does not reconcile.
The second panel agrees on direction, not on magnitude
The other measurement source we use publishes a SendPulse page showing “77.73 % of 76.163 emails… ended up in the recipient’s main mailbox”, with 20.11 percent in spam and 2.17 percent lost. Its test period and methodology are both absent from the page, and its category scores did not survive a second read, so we do not publish those.
Two panels, fourteen points apart on the headline. What they agree on is the number that matters: SendPulse puts somewhere between a fifth and a third of its mail in the spam folder. A healthy platform sits at 2 to 7 percent. That directional agreement, from two independent sources using different methods at different dates, is the safe claim and it is a serious one.
What the vendor says
Nothing. We checked the delivery-rate glossary entry, which offers only a qualitative line about reputation, and the feature pages, which do not reproduce a figure on either read. SendPulse publishes no delivery-rate or inbox-placement number anywhere. We report that as a finding in its own right.
Snapshot
| Legal entity | “SendPulse Inc., 220 E 23rd St #401, New York, NY 10010.” This appears only in the privacy policy. The terms of service name the entity with no address at all. |
|---|---|
| Registration number | Not published anywhere. |
| Founded | “Founded in 2015…” |
| Founders | Not published. No individual is named as a founder on the about page. |
| Scale | “over 200,000 customers”, stated as of 15 July 2026 |
| Footprint | “offices and representatives in the U.S., Brazil, Mexico, and Ukraine” |
| Free plan | 15,000 emails per month to up to 500 subscribers |
| Email plan prices | Do not render. The pricing page serves literal empty placeholders reading “$ /month”, on both reads. The numbers are injected by JavaScript and we could not retrieve them. |
| SMTP prices | These do render: Basic free at 12,000 emails and “50 messages per hour”; Pro $59.88/month for 100,000 emails and “8,000 messages per hour” with a dedicated IP; Enterprise $219.88/month for 500,000 and “20,000 messages per hour” |
| Payment | “We accept Visa and Mastercard credit cards.” |
| Compliance | Nothing published. The /security page returns a 404. SOC 2 and ISO 27001 appear nowhere. GDPR is a self-declaration only, and the privacy policy carries “Updated: 01.04.2024”, well over two years stale. |
The pricing situation is worth pausing on. The same site renders prices on one pricing page and not on another. The SMTP tiers print in full; the email marketing tiers print as empty dollar signs. Third-party listings fill the gap with figures we cannot confirm at source, so we do not publish them.
Eighteen verticals, banned by name
There is no standalone acceptable use policy: the path /legal/aup returns a 404. The list lives inside the terms of service, in a section headed “What content is prohibited”, and it is one of the longest we have read.
| Banned | Escort and dating · pharmaceutical sales · work from home and make money online · online and day trading · MLM · “Affiliate marketing” · debt collection and credit repair · mortgages and loans · supplements · adult novelty · list brokers · illegal goods and pirated software · gambling, betting, poker and casino · political parties · tobacco · alcohol · “Any religious content” · plus sexually explicit messages, hate speech and non-permission lists |
|---|---|
| Not on the list | Cryptocurrency, cannabis, CBD, firearms and forex are not named. We report that as an absence, not as permission. |
Two entries stand out. “Any religious content” banned outright is unusual and worth knowing if you send for a church, a charity or a faith-adjacent publisher. And the ban on “Affiliate marketing” sits oddly beside the review corpus, which contains a reviewer describing themselves as working in affiliate marketing.
Manual moderation is the mechanism, and it recurs across five years
The pattern in the corpus is not mass bans. It is new senders placed under manual review and left without an explanation. A one-star reviewer writing on 30 June 2026:
“My SendPulse account was placed under manual moderation… soon after that my access to SMTP service was blocked. My stated use case was transactional email for my own domain… The support response was vague and did not give me a clear moderation reason.”
A Capterra reviewer in December 2021 raised the same mechanism in fewer words: “manual email moderation slows process”. Two further reviewers, in July and August 2026, describe being billed twice.
Transactional mail for your own domain is the least objectionable use case there is. If that triggers a manual review that is then not explained, the problem is not the moderation, it is the communication around it, and it has been reported the same way from 2021 through to 2026.
The refund policy, which is better than most here
Unusually, SendPulse publishes a dedicated refund page, and the terms are reasonable:
- “Within thirty (30) calendar days of completing your payment… you may ask for a refund of your SendPulse funds which were not spent during this period.”
- Auto-renewal charges can be disputed within fourteen calendar days, and requests are reviewed “within seven (7) business days”
- The first period of a free trial conversion is non-refundable
Refunds are refused where spam complaints “exceeds 0.6-0.9%” or bounces “exceeds 7% within a single mailing list”. Those thresholds are looser than Sender at 0.1 percent, and unlike several competitors here there is a real pro-rata window rather than nothing at all.
SMTP, documented in a code sample and nowhere else
The values are real and we confirmed them twice. Where they live is the problem.
| Setting | Value |
|---|---|
| Hostname | smtp-pulse.com |
| Ports | “You can choose 2525, 465, and 587” |
| SSL | Port 465, shown in the sample as $mail->Port = 465; with SMTPSecure = 'ssl' |
| STARTTLS | Port 587 |
| Hourly rate, Basic (free) | 50 messages per hour |
| Hourly rate, Pro | 8,000 messages per hour |
| Hourly rate, Enterprise | 20,000 messages per hour |
| Authentication | DKIM, SPF and DMARC supported on every tier including free |
| Dedicated IP | Bundled into Pro and Enterprise. A standalone price is not published. |
The hostname and ports appear only inside a PHPMailer code example. Four separate pages titled as SMTP setup guides do not contain them on either read: they tell you to copy the credentials out of your dashboard instead. Two documentation pages that would hold them, smtp-limits and smtp-settings, return 404 while still being linked from the live navigation.
For a platform that sells an SMTP relay as a headline product, that is a documentation failure, and it is the reason the API and MCP score below is not higher.
The 50 messages an hour on the free tier is worth reading twice. It is not 12,000 a month spendable at will: it is a hard hourly gate that makes the free relay unsuitable for anything bursty, such as a signup spike or a password-reset storm.
The API and the MCP server, which are the best-documented parts of the product
The API base is https://api.sendpulse.com, with OAuth2 issuing a bearer token valid for one hour. Rate limits are published as a real table of real values, not as prose:
| Plan | Requests per minute | Requests per day |
|---|---|---|
| Free | 1,000 | 500,000 |
| Standard | 2,000 | 1,000,000 |
| Enterprise | 3,000 | 3,000,000 |
Breaches return 429 Too many requests. Official SDKs cover PHP, Python, Ruby, Java, Node.js and C sharp. A Go client exists but is third party.
There is an official first-party MCP server, at https://mcp.sendpulse.com/mcp, described by the vendor as an “open-source tool”, authenticated with an API key and “Compatible with Claude, OpenAI, Cursor, and other clients”. Open sourcing it is a meaningful choice: it lets you read what the tools actually do before you point an agent at your subscriber list.
The contrast inside this one product is stark. Published rate limits to the request, six SDKs and an open-source MCP server, sitting beside an SMTP relay whose settings you can only find in a PHP snippet and whose own documentation pages 404.
What the reviews say, and why we count two of them as one
SendPulse carries three high ratings: Capterra 4.6 from 773, G2 4.6 from 718, Trustpilot 4.5 from 2,851. That looks like three independent confirmations. It is closer to two.
G2 acquired Capterra, GetApp and Software Advice from Gartner in January 2026, so the two sites are now under one roof and syndicate reviews between them. They return an identical 4.6 on near-identical volumes. We treat them as one corpus, not as two sources agreeing, and we say so wherever the number is used.
We read 25 of the 773 Capterra rows and 20 full Trustpilot rows plus 5 titles. G2 rendered no individual reviews at all, so we read zero of its 718 and quote nothing from it.
Date-bursting in the Capterra sample
Of the 25 rows we read, eleven fall on just three days: five share 14 June 2025, four share 2 April 2026, two share 11 June 2025. All eleven are four or five stars. Every one-star and three-star row in the sample sits on a date of its own.
That is the same shape we found on Selzy, though milder and spread across two years rather than concentrated in one quarter. It is enough that we do not treat the 4.6 as a clean average. Three of the 25 rows, from unrelated industries, independently describe the interface as cluttered, which given how many products are bolted together is not surprising.
The Trustpilot corpus is bimodal, and geographic
The distribution is 77 percent five star, 9 percent four, 3 percent three, 1 percent two, and 10 percent one star. A hollow middle like that, with more one-star than two- and three-star combined, is the signature of two different populations rather than one.
The rows we read point at what separates them. The five-star mass is concentrated in Morocco, Ukraine, Nigeria, Cambodia, the Philippines, Vietnam, Uzbekistan and India, and several are two or three words long. Three of the four rows from the United States are negative. We are reporting a pattern in twenty rows out of 2,851, not a proven property of the whole corpus, but it is a pattern worth knowing before you read 4.5 as a single verdict.
Ratings breakdown
Six categories, each out of 10, judged against the same evidence rules on every platform we cover. The overall is the unweighted arithmetic mean, rounded to one decimal: 6.5 + 7.5 + 4.0 + 5.0 + 6.5 + 7.0 = 36.5, divided by 6, equals 6.1.
Ease of use, 6.5. A Capterra ease sub-score of 4.5 across 773 reviews, and the free tier gets you sending quickly. Held down because three of the 25 rows we read call the interface cluttered, and a product carrying email, chatbots on six networks, SMS, push, a CRM, landing pages and a course builder is inherently more to learn than a mailer.
Automation, 7.5. Automation 360 is a real builder, and the multichannel reach into WhatsApp, Telegram, Instagram, Facebook, Viber and TikTok is wider than anything else we cover. Held below 8 because the depth available on the free tier could not be verified.
Deliverability, 4.0. The driving score in this review. 62.8 percent measured, fourteenth of fifteen, with 33 percent in spam, and a second panel independently putting spam at 20 percent. The vendor publishes no figure of its own to set against either. Held above the floor only because the January 2024 round is now two and a half years old and no newer measurement exists in either direction.
Support, 5.0. Customer service is the lowest of the Capterra sub-scores at 4.4, ten percent of the Trustpilot corpus is one star, and the specific failure reported from 2021 to 2026 is the same one: manual moderation applied without an explanation. Two reviewers also describe double billing.
Pricing value, 6.5. 15,000 emails a month free with a 12,000-email SMTP tier alongside it is genuinely generous, and the refund policy has a real 30-day pro-rata window, which several competitors here do not. Held down hard by the email pricing page not rendering a single price, and by a 50 messages per hour ceiling on the free relay.
API and MCP, 7.0. Rate limits published as an exact table across three tiers, six official SDKs, and an official open-source MCP server. Held down by SMTP settings that exist only inside a code sample, two documentation pages returning 404 from live navigation, and no published compliance certification of any kind.
Who should use SendPulse, and who should not
SendPulse is the right call for
Look elsewhere if
Third-party validation
| Source | Score | Volume | Notes |
|---|---|---|---|
| Capterra | 4.6 / 5 | 773 reviews, 25 read | Ease of use 4.5, customer service 4.4. Product ID 146369, verified to resolve to SendPulse. Discounted: 11 of the 25 rows read fall on three days, all four or five star. |
| G2 | 4.6 / 5 | 718 reviews, 0 read | No individual reviews rendered. Not counted as independent of Capterra: G2 has owned Capterra since January 2026 and the two syndicate reviews, and the scores and volumes are near identical. |
| Trustpilot | 4.5 / 5 | 2,851 reviews, 20 read | 77 / 9 / 3 / 1 / 10 percent by star. Bimodal with a hollow middle. In the rows read, the five-star mass is geographically concentrated and three of four United States rows are negative. |
| Inbox placement | 62.8% | Jan 2024 seedlist round | Fourteenth of fifteen platforms. 33.0 percent spam, 4.2 percent never delivered. Second panel independently reports 20.11 percent spam. |
Top 3 alternatives
| Alternative | Our score | Take it instead if |
|---|---|---|
| Brevo | 7.7 | You want the same marketing-plus-SMTP-plus-SMS shape from the highest-scoring platform we cover, with 88.3 percent measured placement instead of 62.8. |
| MailerLite | 7.4 | Email is the priority and you want 89.8 percent measured, published prices that actually render, and a clean review corpus. |
| SMTP2GO | 7.5 | You came for the relay. Eight ports, a hostname per region, no hourly gate on the free tier, and the highest sub-scores on this site. |
Continue reading
- SendPulse pricing, the free tier and the SMTP plans that do render
- SendPulse API, published rate limits and the open-source MCP server
- SendPulse integrations
- Brevo review
- MailerLite review
- SMTP2GO review
Frequently asked questions
Is SendPulse free?
Yes. The free email plan covers 15,000 emails a month to up to 500 subscribers, and there is a separate free SMTP tier of 12,000 emails capped at 50 messages per hour. Chatbots, web push and the CRM are also reachable at no cost. The hourly ceiling on the free relay is the limit most people hit first, because it makes any burst of mail queue rather than send.
How much does SendPulse cost?
We can tell you the SMTP prices and not the email marketing prices. The SMTP tiers render in full: Pro at $59.88 a month for 100,000 emails with a dedicated IP, Enterprise at $219.88 a month for 500,000. The email marketing pricing page serves empty placeholders reading dollar sign slash month on both of our reads, because the figures are injected by JavaScript. Third-party listings publish numbers we could not confirm at source, so we do not repeat them. Payment is by Visa or Mastercard.
What are the SendPulse SMTP settings?
The hostname is smtp-pulse.com and the ports are 2525, 465 and 587, with 465 for SSL and 587 for STARTTLS. Sending is capped by the hour rather than by the day: 50 messages an hour on the free Basic tier, 8,000 on Pro and 20,000 on Enterprise. DKIM, SPF and DMARC are supported on every tier including free. Be aware that these values are published only inside a PHPMailer code sample; the pages titled as SMTP setup guides do not contain them, and two related documentation pages return 404.
Is SendPulse good for deliverability?
On the only independent evidence that exists, no. The January 2024 seedlist round, the last one ever run, measured SendPulse at 62.8 percent inbox placement with 33 percent going to spam, which was fourteenth out of fifteen platforms tested. A second independent panel separately reports 20.11 percent of its mail in spam. A healthy platform sits at 2 to 7 percent. SendPulse itself publishes no delivery-rate figure anywhere on its site, so there is no vendor number to weigh against those two.
Can my SendPulse account be blocked?
Yes, and the usual mechanism is manual moderation of new senders rather than an outright ban. Eighteen verticals are prohibited by name in the terms, including affiliate marketing, gambling, MLM, loans, supplements, tobacco, alcohol, political parties and any religious content. Reviewers from 2021 through to June 2026 describe accounts placed under manual moderation and SMTP access blocked, with the recurring complaint being that no clear reason was given rather than that the review happened at all.
Does SendPulse have an MCP server?
Yes, an official first-party one at mcp.sendpulse.com/mcp, which the vendor describes as an open-source tool authenticated with an API key and compatible with Claude, OpenAI, Cursor and other clients. That places SendPulse in the minority of platforms we cover that ship one, and open sourcing it means you can read what the tools do before pointing an agent at your subscriber list.
Changelog
| Date | Change |
|---|---|
| 27 August 2026 | First publication. The 62.8 percent figure was confirmed on a second literal read of the source article, together with its methodology wording and its rank of fourteenth of fifteen. A reconciliation defect was found in the source table, where the SendPulse breakdown sums to 106.8 percent against about 100 for every other row; the headline, spam and missing figures are published and the corrupt split is not. The panel hub page returned contradictory readings across two passes and was discarded in favour of the dated article. 25 of 773 Capterra rows read with date-bursting identified; 20 of 2,851 Trustpilot rows read; G2 rendered none of its 718 and is neither quoted nor counted as independent of Capterra, G2 having owned Capterra since January 2026. Email plan prices do not render and are recorded as not published rather than taken from third parties. SMTP hostname, ports and hourly rates were confirmed twice, and the fact that they appear only inside a code sample is reported. Search demand was measured before the FAQ was written. |
This review follows our email infrastructure testing methodology. We disclose affiliate relationships in our editorial independence policy.
