How Resend Became the Default Email Layer for the AI Coding Era

Case Study · Developer Tooling & AI

Company: Resend · Founded: January 2023 · Backer: Y Combinator W23 · Category: Email API / Transactional

• 100k+ developers on the platform within 15 months of launch
• 95 Y Combinator companies using Resend — #1 deal in the W23 batch
• 5 lines of code to send your first email — the fewest of any major API

01 — Background: Email Infrastructure, Frozen in 2010

When Zeno Rocha was Chief Product Officer at Liferay and later VP of Developer Experience at WorkOS, he kept running into the same problem: transactional emails were vanishing into spam folders, and the tools available to fix it — SendGrid, Mailgun, Mailchimp — felt like they had been built for a different era of software development. Documentation was sprawling and inconsistent. APIs were designed around marketing workflows, not developer primitives. Setting up a simple password-reset email involved wrestling with arcane HTML templates, unreliable deliverability, and zero visibility into what happened after the send.

Rocha's insight was simple but sharp: email is one of the first things every software product needs, which makes it one of the highest-leverage places to compete on developer experience. "Email is the most underrated technology in the world," he said. "It's been around forever and it's still the most effective way to communicate with your users." What it lacked was a Stripe-like moment — a complete rethinking of how the API should feel to build on.

"Paul Graham described Resend as 'the Stripe for Email.' That's the bar we set for ourselves." — Zeno Rocha, Founder & CEO, Resend

02 — Origin: An Open-Source Hook, Then a Company

Before Resend existed as a company, Rocha and co-founder Bu Kinoshita launched React Email — an open-source project that let developers write email templates as React components rather than archaic HTML tables. The response was immediate. React Email reached over 12,000 GitHub stars and made clear that the developer community was starving for a modern take on email primitives.

• Late 2022 — React Email launches. Open-source library for writing emails in React. Tens of thousands of developers immediately adopt it. GitHub stars climb quickly.
• January 2023 — Resend incorporated; YC W23 batch. Rocha quits his job and pays $20,000 of his own money for the Resend.com domain — before the YC money arrived. Acceptance into Y Combinator provides validation and a built-in distribution channel.
• Mid-2023 — Public launch & product-market fit. 6,469 developers on the waitlist before public launch. Within weeks, Resend is the #1 most-used deal from the YC W23 batch, used by 95 YC companies.
• Early 2024 — 100,000 users; Series A. Fifteen months after officially starting, Resend reaches 100,000 developers. A Series A round follows to fund scaling infrastructure and the team.
• February 2025 — new.email & LLM-native positioning. Resend launches new.email — an AI-powered email builder — and publishes llms.txt and llms-full.txt, establishing a machine-readable source of truth for every AI coding tool in the ecosystem.

03 — The AI Coding Wave: Why Vibe-Coding Tools Reach for Resend

Between 2023 and 2025, a new class of development tool exploded into mainstream use. Platforms like Lovable, Bolt.new, v0 by Vercel, Replit Agent, and Cursor shifted millions of builders — many of them non-technical — from writing code to describing what they wanted in plain English. These tools generate entire full-stack applications from a prompt. And every application they generate eventually needs to send an email.

When an AI coding tool writes email-sending code, it defaults to the provider it knows best — the one whose API pattern is clearest, whose documentation is most structured, and whose code examples are most likely to be correct on the first try. Resend won that competition decisively, for reasons that compound on each other:

• Five-line hello world — Resend's send API uses a single async function call with a destructured { data, error } response — a pattern AI models generate reliably. Competitors require dozens of lines before the first email goes out.
• React Email synergy — Because React Email is Resend's open-source sister project, the two are naturally paired in documentation, tutorials, and community examples. AI tools trained on that corpus learn the pairing.
• llms.txt — docs for AI — Resend publishes a machine-readable llms.txt and llms-full.txt — a structured summary of its entire API designed specifically for LLM context windows, not human browsers.
• Consistent, camelCase API — Every parameter is camelCase. Every error is handled the same way. There are no legacy endpoint variants to confuse a model. The API has a single right answer for every question.
• MCP Server support — Resend ships an official Model Context Protocol server, letting AI agents like Claude Desktop, Cursor, and others send emails via natural language — no custom integration code required.
• Quick onboarding — A free tier, a test mode that never sends real emails, and a single environment variable for the API key means AI-generated apps reach "first successful send" without friction that would prompt retries.

04 — The Code Advantage: What the AI Generates

When an AI coding tool is asked to "add email notifications," the pattern it reaches for looks like this. The simplicity is not accidental — Resend designed its SDK for exactly this experience, and that decision made it the path of least resistance for every LLM trained on developer tutorials.

Contrast this with Amazon SES, which requires IAM role configuration, multi-page permission setup, and sandbox approval before a single email can be sent. Or SendGrid's v3 API, where the personalization model introduces concepts like "personalizations arrays" for what developers expect to be a simple send. From the perspective of an LLM generating code, Resend's API has exactly one correct answer to every common email question — and that predictability is gold.

"As an agent scanning documentation, SendGrid's sheer volume of pages with overlapping content creates ambiguity about which is the canonical source." — Courier Blog, Best Email API for Developers 2026

05 — The Documentation Strategy: Docs as Product, Product as Distribution

Resend's documentation philosophy, articulated by Rocha early on, is that documentation is not adjacent to the product — it is part of the product. The company invested heavily in Mintlify-powered docs that are fast, well-structured, and aesthetically consistent with the brand. The result was documentation that developers shared on inspiration channels and that became a benchmark others tried to emulate.

But the more strategically significant move was publishing llms.txt and llms-full.txt — dedicated, machine-readable versions of the documentation formatted for LLM context windows. These files contain structured guardrails for AI-generated code: verification steps an AI model should run before returning any Resend-related solution, canonical parameter names, and framework-specific guides. The implication is clear: Resend is not just optimizing for human developers finding their docs on Google. They are optimizing for AI tools being the first place a developer goes.

They also built out an explicit MCP (Model Context Protocol) server, meaning Claude Desktop, Cursor, and other AI-native IDEs can connect to Resend as a native tool — letting agents send emails through natural language commands rather than generated code.

06 — Ecosystem Fit: Built-In to the AI Coding Stack

The AI coding tools that reached mass adoption in 2024 and 2025 share a common DNA: they connect to Supabase for the database, Stripe for payments, and — increasingly — Resend for email. The pattern is not coincidental. Each of these tools occupies a "developer-first" position in its category, with clean APIs and opinionated defaults that make them easy for AI to generate correctly.

Resend's adoption spans across Lovable, Bolt.new, v0 by Vercel, Cursor, Replit Agent, Claude Code, and Claude Desktop (MCP).

The company also saw firsthand how AI tools were changing who builds software. Lovable, Bolt, and v0 were generating working applications for non-technical founders and product managers. That cohort was newly capable of building software — but had never set up an email provider before. The path of least resistance, for the AI writing their code and for them, was Resend.

"We've seen firsthand how products like Lovable, Bolt, and v0 are changing the definition of a developer." — Resend blog, Introducing new.email (February 2025)

07 — Lessons: What Resend Got Right

1. API design is marketing. Every design decision that made Resend's API simpler — camelCase params, the { data, error } pattern, one correct path for every operation — lowered the failure rate for AI-generated code. A lower failure rate means more developers who reach "it works" on the first try, and more apps deployed with Resend in the stack.

2. Open source earns the right to sell. React Email created tens of thousands of developers who already trusted the Resend team's judgment before the paid product launched. The GitHub stars were not just vanity — they were a distribution channel and a trust signal.

3. Document for your actual reader. In 2025, that reader is increasingly an LLM. Publishing llms.txt and llms-full.txt was a deliberate bet that AI-generated code would become a primary source of new customer acquisition — and that bet has compounded as vibe-coding tools went mainstream.

4. Find the "Stripe moment" in an unsexy category. Email infrastructure is not glamorous. But Stripe proved that re-imagining the developer experience for a commodity service can produce a category-defining company. Resend applied the same thesis to email, in a year when the audience of capable builders was expanding faster than anyone expected.

5. MCP is the new SDK. By building a first-party Model Context Protocol server, Resend positioned itself not just as a library developers import but as a tool that AI agents use directly — a meaningful shift in how software gets assembled.