# Which transactional email service do AI models recommend? Source: Memetik Research, https://www.memetik.ai/research/transactional-email-according-to-ai. Published 2026-09-27. Data: Memetik Index 2026-09. By [Memetik](https://www.memetik.ai/research). Resend was named in all 50 recorded AI answers and first in 28. See every service's count, the gap to first place and where 10 models disagree. ## Key findings 1. All 10 models put Resend at the top of their own counts. The leader stays the same whichever assistant a buyer asks. 2. Amazon SES, SendGrid and Mailgun reach most answers and almost never open one. Answers usually offer them after a first pick. 3. mailtrap.io is the most-cited site in the answers, yet the three OpenAI models never name Mailtrap. 4. The guides Google ranks build a different shortlist. Postmark, SendGrid and Mailgun appear in every captured guide. Resend appears only in the Loops guide. 5. Loops, Mailtrap and Userpilot publish guides that list their own products. Read each ranking as its publisher's view. ## Chart: Transactional email, edition 2026-09 | Vendor | Answer share | Named first | | --- | --- | --- | | Resend | 100% | 56% | | Amazon SES | 96% | 4% | | Postmark | 94% | 30% | | SendGrid | 86% | 2% | | Mailgun | 78% | 0% | | Mailtrap | 52% | 8% | | Brevo | 26% | 0% | | Loops | 10% | 0% | 50 answers. Edition source: https://www.memetik.ai/data/2026-09/transactional-email.json Resend. 50 of 50 recorded AI answers named [Resend](/vendors/resend), and 28 named it first. Postmark is the only other service that opens answers often. Amazon SES appears almost as often as Resend but rarely comes first. The counts come from the 2026-09 edition of the Memetik Index. ## Which transactional email services did the models name? The models named 12 of the 14 tracked services. Resend leads on names and on first place. Postmark is second on first place and sits close to Resend on average position. | Vendor | Named | Answer share | Named first | Named-first share | Avg position | | --- | --- | --- | --- | --- | --- | | Resend | 50/50 | 100% | 28/50 | 56% | 1.84 | | Amazon SES | 48/50 | 96% | 2/50 | 4% | 3.4 | | Postmark | 47/50 | 94% | 15/50 | 30% | 1.91 | | SendGrid | 43/50 | 86% | 1/50 | 2% | 4.16 | | Mailgun | 39/50 | 78% | 0/50 | 0% | 4.44 | | Mailtrap | 26/50 | 52% | 4/50 | 8% | 4.15 | | Brevo | 13/50 | 26% | 0/50 | 0% | 6.69 | | Loops | 5/50 | 10% | 0/50 | 0% | 4.6 | | Knock | 5/50 | 10% | 0/50 | 0% | 4.6 | | SparkPost | 2/50 | 4% | 0/50 | 0% | 6 | | Mailjet | 1/50 | 2% | 0/50 | 0% | 7 | | Plunk | 1/50 | 2% | 0/50 | 0% | 7 | Two services carry the top of the answers. Resend is named everywhere. [Postmark](/vendors/postmark) is named slightly less often than [Amazon SES](/vendors/amazon-ses) and opens far more answers. Amazon SES, [SendGrid](/vendors/sendgrid) and Mailgun fill the middle. They appear in most answers and open almost none. Below Mailgun the counts thin out. Brevo, Loops, Knock, SparkPost, Mailjet and Plunk each reach a minority of answers and open none. The models' own words show what a name carries. Gemini, asked what a developer should use, wrote: > If developer experience is your top priority, Resend is the current favorite. A name is not always praise. Claude Opus 5 put Resend at the top of one answer and added: > Weakness: younger company, less mature deliverability tooling for high-volume senders. Both answers count as naming Resend. ## How big is the gap between being named and being named first? Amazon SES has the widest gap, followed by SendGrid and Mailgun. All three are named in most answers and open almost none of them. | Vendor | Gap (points) | | --- | --- | | Amazon SES | 92 | | SendGrid | 84 | | Mailgun | 78 | | Postmark | 64 | | Resend | 44 | | Mailtrap | 44 | | Brevo | 26 | | Loops | 10 | | Knock | 10 | | SparkPost | 4 | | Mailjet | 2 | | Plunk | 2 | A small gap at the bottom is a small share meeting zero. Mailjet and Plunk have small gaps because they were barely named. Equal gaps can mean different things. Resend and Mailtrap sit level. Resend's gap sits on top of a full answer share. Mailtrap's sits on a much smaller one. Among the most-named services, only Resend has a smaller gap than Postmark. The answers explain the Amazon SES gap themselves. Several frame it as the pick for AWS teams or large volumes. Claude described it this way: > Amazon SES has the highest ceiling, with production-access requests able to get you to thousands of emails per second, but it has the rawest API, meaning you'll write your own retry and batch logic. For a buyer, that makes Amazon SES a conditional option. Answers offer it for a situation, such as cost or an AWS stack, and another service usually leads. ## Do the models agree with each other? On the leader, yes. Each of the 10 models ranks Resend first on its own counts, and the provider families agree. Below Resend the models split. | Vendor | GPT-5.6 Sol | ChatGPT | GPT-5.6 Luna | Claude Opus 5 | Claude | Claude Fable 5 | Gemini | Gemini 3.5 Flash | Perplexity | Sonar Reasoning Pro | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Resend | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | | Amazon SES | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 4/5 | 4/5 | | Postmark | 4/5 | 3/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | 5/5 | | SendGrid | 4/5 | 5/5 | 5/5 | 4/5 | 4/5 | 5/5 | 4/5 | 2/5 | 5/5 | 5/5 | | Mailgun | 3/5 | 3/5 | 5/5 | 5/5 | 5/5 | 4/5 | 3/5 | 4/5 | 3/5 | 4/5 | | Mailtrap | 0/5 | 0/5 | 0/5 | 3/5 | 4/5 | 3/5 | 4/5 | 4/5 | 3/5 | 5/5 | | Brevo | 0/5 | 0/5 | 1/5 | 2/5 | 4/5 | 2/5 | 1/5 | 0/5 | 1/5 | 2/5 | | Loops | 0/5 | 0/5 | 0/5 | 2/5 | 0/5 | 0/5 | 1/5 | 2/5 | 0/5 | 0/5 | | Knock | 0/5 | 0/5 | 0/5 | 3/5 | 1/5 | 1/5 | 0/5 | 0/5 | 0/5 | 0/5 | | SparkPost | 1/5 | 0/5 | 1/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | | Mailjet | 0/5 | 0/5 | 0/5 | 0/5 | 1/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | | Plunk | 0/5 | 0/5 | 0/5 | 1/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | 0/5 | Mailtrap has the widest split of any service. Its pattern has its own section below. Brevo is the next sharpest. Claude names it in most of its answers. GPT-5.6 Sol, ChatGPT and Gemini 3.5 Flash never name it. Knock appears only in Anthropic answers: Claude Opus 5, Claude and Claude Fable 5. Loops appears only in answers from Claude Opus 5, Gemini and Gemini 3.5 Flash. SparkPost appears only in GPT-5.6 Sol and GPT-5.6 Luna. The top names are steadier. ChatGPT names Postmark least often of the 10 models. Gemini 3.5 Flash names SendGrid least often. For a buyer, the assistant changes the tail of the shortlist. The top of it stays put. ## Which sources do the answers cite? mailtrap.io is the most-cited host in the recorded answers. It is also the only top host that belongs to a tracked vendor. | Host | Citations | | --- | --- | | mailtrap.io | 193 | | nitrosend.com | 112 | | agentmail.to | 112 | | smtp2go.com | 104 | | sequenzy.com | 85 | | reddit.com | 58 | | maestra.io | 37 | | ventureharbour.com | 37 | The other vendor-owned hosts sit far lower: postmarkapp.com at 22, loops.so at 16, mailgun.com at 15 and aws.amazon.com at 14. Resend's own site is cited less often than any of them. Several top hosts serve comparison pages. The answers cite best-of roundups on smtp2go.com and maestra.io, for example. reddit.com is the one forum among the top hosts. Citation counts include the citations returned with the recorded answers. Coverage varies by model. Gemini citation redirects are resolved to real hosts before counting. ## Does the most-cited site get its vendor named? Not by every model. The Memetik Index count puts mailtrap.io ahead of every other cited host, yet Mailtrap trails Resend, Amazon SES, Postmark, SendGrid and Mailgun on names. The split runs along model lines. GPT-5.6 Sol, ChatGPT and GPT-5.6 Luna never name [Mailtrap](/vendors/mailtrap). Sonar Reasoning Pro names it in every answer. Claude, Gemini and Gemini 3.5 Flash name it in most of theirs. When Mailtrap is named, it opens more answers than Amazon SES or SendGrid, both of which are named far more often. Mailtrap's own guide is also the first organic result in the captured Google search for this question. So Mailtrap leads the source count and holds the top organic result, yet no OpenAI answer names it. Being cited and being named are separate outcomes. For a vendor, citations are an input worth tracking. Names are the outcome a buyer reads. For a buyer, the assistant you ask decides whether Mailtrap reaches your shortlist at all. ## How was the sample built? 10 models x 5 fixed prompts = 50 recorded answers, in the 2026-09 edition. The five recorded questions: - What is the best transactional email API for a developer? Name specific products. - Which transactional email API would you recommend to a developer in 2026? - Compare the top transactional email API options right now. - I'm a developer and I need a transactional email API. What should I use and why? - Best transactional email API for a SaaS sending signup and receipt emails? | Family | Models | Answers | | --- | --- | --- | | OpenAI | GPT-5.6 Sol, ChatGPT, GPT-5.6 Luna | 15 | | Anthropic | Claude Opus 5, Claude, Claude Fable 5 | 15 | | Google | Gemini, Gemini 3.5 Flash | 10 | | Perplexity | Perplexity, Sonar Reasoning Pro | 10 | Each model carries equal weight. A vendor is named when its name or a known alias appears in the answer text. Answer share is the share of the 50 answers that name the vendor. Named first means the vendor was the earliest tracked vendor in the answer. Average position is where the vendor falls among tracked names in the answers that name it, averaged; lower is earlier. The gap is answer share minus named-first share, in points. Every raw answer sits in the [Index entry](/index/transactional-email), and the counts are in the [edition data file](/data/2026-09/transactional-email.json). The [method](/method) sets out the full counting rules. ## How does this sit against the transactional email guides? The Memetik Index counts the names AI answers produce. The guides that rank in Google for this question rank products for purchase. Loops, Mailtrap and Userpilot list their own products in theirs. | Guide | First listed | Lists itself | Lists Resend | | --- | --- | --- | --- | | Zapier | SendGrid | No | No | | Userpilot | Postmark | Yes | No | | Mailtrap | Mailtrap | Yes | No | | Loops | Loops | Yes | Yes | Zapier's guide lists SendGrid, Postmark, Mailgun, MailerSend, Mailmodo, Waypoint and Zoho ZeptoMail, in that order. Zapier discloses that it may earn a commission when readers click some of its links. Its author left out tools that require coding to customise all emails. Amazon SES appears only as the infrastructure under Waypoint. Userpilot's guide is written by a Userpilot product manager. It puts Postmark first and Userpilot second, then Mailchimp, Amazon SES, SendGrid, Mailgun, HubSpot, Mailtrap and Brevo. It closes with an invitation to book a Userpilot demo. Mailtrap's guide tested Mailtrap, SendGrid, Mailgun, Brevo, Amazon SES and Postmark. Its disclaimer says Mailtrap publishes the comparison and ranks its own product. It lists Mailtrap first and recommends Mailtrap for most SaaS and product teams. The Loops guide states that Loops is its product and that it appears first. Postmark, Resend, SendGrid, Amazon SES, Mailgun, Brevo and Mailchimp Transactional follow. The two shortlists overlap only in part. Postmark, SendGrid and Mailgun appear in every captured guide. Resend, the name at the top of the AI answers, appears only in the Loops guide. Framing may explain part of the gap. Zapier asks for drag-and-drop design. All five recorded prompts ask for an API. One model raised the same point in its answer. Claude Opus 5 opened a comparison with: > One caveat up front: many "best of" roundups are published by vendors who rank themselves in the list, so I've flagged that where relevant. Use the guides for features and prices, read as each publisher's view. Use the counts to see what the assistants say when a developer asks. ## What can these counts not tell you? They cannot tell you which service is better. The counts record names in AI answers. Product quality, deliverability and price sit outside them. Each model answered each prompt once. The counts are a descriptive snapshot and do not estimate repeatability. They cover one dated edition, 2026-09. Alias matching is string-based. The answers came through model APIs, which can differ from consumer chat apps. The prompts are in English. ## Frequently asked questions ### Which email service is best for transactional emails? No single service suits every sender, and the sources disagree. In the recorded AI answers, Resend is named most and opens the most answers, with Postmark next on first place. All four captured guides list Postmark, SendGrid and Mailgun. Loops, Mailtrap and Userpilot rank their own products in their guides. The Loops guide suggests running the same password-reset and receipt flow through your final two candidates before you commit. ### Why are Customer.io and Courier missing from the tables? No model named them. Both are tracked in the transactional email category, and neither appeared in any recorded answer. A tracked vendor with no mentions gets no row. ### Can a vendor pay to be named in the Memetik Index? No. No vendor pays. The counts come only from the recorded answers to fixed prompts, and every answer is published. ### Will the counts change in the next edition? They can. The five prompts stay fixed between editions, so the next edition asks the same questions and any change shows up against this one. Past and future editions sit in the [research archive](/research). ## Method and limits Edition 2026-09: 5 fixed buyer prompts per category, sent to 10 models through their APIs. Each category contains 50 recorded answers. Answer share is the fraction of answers naming a vendor. Named first records the vendor's position before other tracked vendors, not the strength of a recommendation. This is one dated sample; small differences can reflect answer variation. API responses can differ from consumer chat products. Full method: https://www.memetik.ai/method. Memetik is a Searchmaxxed publication. Charts and figures: CC BY 4.0.