← Writing

25 Sept 2026 · 5 min read

Resend vs SendGrid: which one a small product should pick in 2026

We send every MeltsWell email through Resend: 13 transactional emails, one file of wiring. Here's how it compares to SendGrid on price, free tier, developer experience and deliverability, and when SendGrid is still the right call.

Short answer: for a new product sending transactional email (sign-in codes, receipts, "your page is live"), pick Resend. For high-volume marketing email with a team of non-developers running campaigns, SendGrid still has more of the tooling.

We use Resend for MeltsWell, our marketplace for home bakers, and have since the second day of the project. Here's what that looks like in practice, and where SendGrid wins.

Resend vs SendGrid at a glance

ResendSendGrid
Free tier3,000 emails a month, 100 a day, forever60-day trial, 100 a day, then paid
First paid plan$20/month for 50,000 emails$19.95/month (Essentials) for 50,000
Dedicated IP$30/month add-on on Scale plansIncluded from Pro ($89.95/month)
TemplatesReact Email, or plain HTMLDynamic templates with a visual editor
Marketing emailBroadcasts, priced by contacts (1,000 free)Marketing Campaigns, separate plans
Setup time for a developerMinutesAn hour or so, more with the dashboard
CompanyIndependent, launched 2023Part of Twilio, around since 2009

Prices from each pricing page in September 2026.

The price difference at 50,000 emails is five cents. The difference that matters for a new product is the free tier. SendGrid dropped its permanent free plan in 2025. Resend's free plan covers a product with a few hundred active users, which is most MVPs.

What we actually send through Resend

MeltsWell sends 13 different emails. None of them is marketing.

To creators: you're live, approved but waiting on a plan, not approved (with the reason), trial ending, payment failed, moved to the Free plan, and the code to confirm account deletion.

To me, as the admin: sign-in codes for the admin panel, a second code before deleting a creator, a new page waiting for review, a creator deleted their account, a payment that arrived after deletion, and a possible duplicate payment.

The whole integration is one file. It creates the client and calls resend.emails.send() with a from address, a to address, a subject and HTML. Every email above goes through that one function. No SDK wrappers, no queue, no template IDs to keep in sync with a dashboard.

The emails themselves are plain HTML strings with inline styles, written in TypeScript next to the code that sends them. Billing emails switch between French and English based on the creator's country. We could use React Email and probably will once there's a designer involved. For 13 simple emails, a function that returns a string is easier to change.

Where Resend is better

The developer experience. The API has about as many concepts as you'd guess: domains, emails, and broadcasts if you want them. You add a domain, paste three DNS records, send. The dashboard shows every email with its status and full content, which is how we found out why our test emails were bouncing.

Templates live in your code. With SendGrid's dynamic templates, the email lives in their dashboard and your code refers to it by ID. That's good if a marketer edits emails. It's bad if you want the email reviewed in the same pull request as the feature that sends it.

The free tier is real. A new product can run on it for months.

React Email. Resend's team also makes React Email, and the two fit together. If your app is already React, your emails can use the same components.

Where SendGrid is better

Marketing tooling. Segments, a drag-and-drop editor, A/B tests on subject lines, scheduled campaigns. Resend has Broadcasts, which covers a founder sending a monthly update. SendGrid covers a marketing team.

Dedicated IPs at volume. If you're sending hundreds of thousands of emails a month and want your own sending reputation, SendGrid's Pro plan includes an IP. On Resend it's an add-on.

Age. SendGrid has been around since 2009 and most tools that integrate with "an email provider" support it. If a third-party tool in your stack needs to send mail, check it supports Resend before you commit.

Deliverability

Both send from good infrastructure, and both land in the inbox if you set up your domain properly. That means SPF, DKIM, and a DMARC record, and sending from a subdomain so a mistake can't damage your main domain's reputation. MeltsWell sends from hello@mail.meltswell.com for that reason.

What hurts deliverability on a small product is what you send, not who sends it. Transactional email that people expect (a code they just asked for) rarely has a problem on either.

Three things we learned the hard way

Test addresses bounce. When we tested the mobile app's account deletion flow with example.com addresses, Resend rejected them. Use your provider's test addresses, or a real inbox you control. Resend offers @resend.dev addresses for this.

Your email provider is not your scheduler. Trial reminders go out from a daily job. We run those as GitHub Actions hitting an API route with a secret, not inside Resend. Whatever provider you pick, the "when to send" logic belongs in your app.

Deduplicate in your database. A cron that runs twice sends twice. We store a trial_reminder_sent_at column and check it before sending. Neither provider protects you from your own retries unless you pass an idempotency key.

Which one to pick

  • New product, developers own email, mostly transactional: Resend.
  • You'll send under 3,000 emails a month for a while: Resend, free.
  • A marketing team will build and send campaigns without a developer: SendGrid, or a dedicated marketing tool next to Resend.
  • Hundreds of thousands a month and you need a dedicated IP included: SendGrid Pro.
  • You're already on SendGrid and it works: stay. Switching providers is a day of work and buys you nothing if nothing is broken.

Email is one line in the budget of an app, and the choice matters less than getting the DNS right. If you're choosing the rest of the stack too, what an app costs and what goes in an MVP cover the bigger decisions. And if you'd rather someone else made them, tell us what you're building.

Written by Badr El Adraoui, founder of Node & Scale. Building something? Build with us.