Loops in 2026: CLI, AI Skills, and Conversion Tracking Arrive

A developer building a SaaS product hits the same wall every time email comes up. The marketing team wants a drag-and-drop campaign builder. The engineers want to fire a welcome email from a signup handler with three lines of code. Most email platforms serve one crowd and treat the other as a second-class citizen, so companies end up paying for two tools that never share a contact list. Loops has spent 2026 attacking that split from the developer side, and the releases it shipped this year say a lot about who it wants using the product.

The headline change landed on April 16, 2026, when Loops released its first command-line tool, an agent Skills system, and a feature called Arrays. A few weeks later it turned on Goals, which finally lets the platform track whether an email drove a conversion instead of just an open. Taken together, these moves push Loops further toward being the email layer you drive from code, not from a dashboard.

Loops has always pitched itself as “the email platform built for SaaS,” combining marketing and transactional email under one contact list. The 2026 releases are the clearest sign yet of what that positioning means in practice: less clicking, more scripting, and an assumption that an AI agent might be doing the work.

The CLI Makes Loops Drivable From the Terminal

Loops now ships a command-line tool that reaches parity with its existing API. You can send a transactional email, update a contact, or manage a list without opening the app, straight from your terminal or a CI pipeline. The tool is open source on GitHub, which matters for teams that want to audit exactly what runs against their sending infrastructure.

The reason this is more than a convenience feature: it makes Loops scriptable in the same places your other developer tooling already lives. A deploy script can send a release note. A scheduled job can clean a mailing list. A local test can fire a real transactional template to your own inbox before you ship. None of that required a CLI before, but doing it through raw API calls meant writing and maintaining boilerplate that most small teams never got around to.

The honest limitation is that a CLI is a starting point, not a destination. Loops itself calls this “the first step in making Loops drivable via API,” and the tool has parity with the current API rather than exceeding it. If a workflow was impossible through the API before, the CLI does not add it. What it changes is friction, not capability.

Skills Aim Squarely at AI Agents

The more forward-looking release is Skills. Loops published a set of installable Skills that teach an AI coding agent how to work with Loops correctly, and you can install all of them or pick individual ones from the Skills page. There are three at launch: an API skill for working with contacts, lists, events, and transactional email; a CLI skill for running Loops from the terminal; and a best-practices skill covering deliverability and message-type decisions.

The practical effect is that you can tell your agent to “wire up Loops in this app and trigger a welcome email when a user signs up,” and it will reach for the SDK, the HTTP API, or the CLI depending on where the code lives. Without a skill, an agent guesses at your integration from stale training data and frequently gets the SDK method names or the event payload shape wrong. With one, it works from Loops’ current documented patterns.

The best-practices skill is the quietly useful one. It can be installed independently of Loops, and it encodes judgment an agent usually lacks: when a message should be transactional versus marketing, how to avoid deliverability traps, and how to structure an email so it does not trip spam filters. That is the kind of context that separates a working integration from one that lands your first thousand emails in spam.

The weakness here is scope. Skills only help if you are already committed to Loops and already working through an AI agent. If your team writes integrations by hand or has no interest in agent-driven development, Skills are dead weight. They also cannot fix a bad decision — an agent that decides to blast a marketing email to your transactional list will still do damage, faster.

Goals Finally Close the Conversion Loop

For most of its life, Loops could tell you an email was opened and clicked but not whether it did anything useful. Goals, which moved from alpha to live in mid-2026, changes that. Loops now tracks impressions, enrollments, and conversions, so you can measure activation, upgrades, demo bookings, purchases, retention, or any contact state that signals the send actually worked.

This is a meaningful gap to close for a SaaS-focused tool. A welcome sequence that gets a 60% open rate looks great until you learn it drove zero activations. Conversion tracking lets you tie an email to the outcome you care about — a user reaching the “aha” moment, a free account upgrading, a churned customer coming back — rather than to a vanity metric. For product-led companies, that connection is the entire point of lifecycle email.

Loops has also signposted what comes next: a Content API for Campaigns, still in the “shipping soon” column at the time of the April release, that will let you create and edit campaigns over the API and use Claude or Codex to update email content programmatically. API support for the Loop Builder and transactional emails is slated to follow. If Loops delivers those, the dashboard becomes optional for far more of the workflow than it is today.

Arrays and the Smaller Fixes

Arrays is the least flashy release and one of the most practical for transactional senders. It lets you pass a variable number of items in a single payload — line items on an invoice, products in an order confirmation, entries in a digest — without building a separate template for every possible length. Anyone who has hacked around a fixed-length template to render a five-item receipt will recognize the problem it solves.

Alongside these, Loops shipped new Stripe subscription events, a suppression-list endpoint for managing addresses that should never receive mail, zoom controls in the Loop Builder, and a refactored Components feature. Individually minor, together they reflect a product being steadily hardened for teams that send real transactional volume rather than the occasional newsletter.

Pricing Reality

Loops keeps its pricing unusually simple: one plan, all features included, priced by contact count. There are no feature tiers to decode and no per-email charges. Team seats are unlimited on every plan, which is rare and genuinely useful for a growing company.

The free plan covers up to 1,000 contacts and 4,000 sends per month with every feature available, but it stamps a “Powered by Loops” footer on your emails. Paid plans start at $49/month for up to 5,000 subscribers on the Starter tier, with unlimited sends and the branding removed. Higher tiers run from roughly $99 to $249 as your list grows, and a Scale plan sits around $399. Annual billing knocks off a chunk versus paying monthly. These figures come from the Loops pricing page; check it directly, since contact-based pricing means your exact rate depends on your list size.

The catch is the entry point. That $49 first paid tier is steep compared to tools that start at $9 or $10, and there is no cheap paid step between free and Starter. If you have 1,200 contacts and want to remove the branding, you jump straight to the $49 plan even though you are barely over the free limit. Loops is priced for funded SaaS companies, not side projects. For a fuller breakdown, see the Loops pricing page on this site.

Where Loops Still Falls Short

The 2026 releases sharpen Loops’ developer story but do not fix its structural gaps.

Marketing features are thin next to dedicated tools. Loops handles broadcasts and sequences well, but if your growth team wants a deep visual automation canvas, a large template gallery, or advanced landing-page and form builders, it will feel limited compared to a tool like MailerLite or Kit. Loops trades breadth for a clean, code-first core.

It is built for SaaS, and it shows. The tagline is not marketing spin — the product assumes a product-led company with an app that emits events. If you run an ecommerce store, a content newsletter, or a service business without an app firing signals into Loops, much of its value goes unused, and you would be better served elsewhere.

The developer bet excludes non-technical teams. A CLI, Skills, and an API-first roadmap are worthless to a solo marketer who wants to log in and click. Everything Loops shipped this year rewards teams with engineers in the loop. That is a deliberate choice, but it narrows who should adopt it.

The API is still catching up to the UI. Campaigns, the Loop Builder, and transactional emails are only partway to full API coverage. The vision of a fully drivable platform is real but unfinished, and today you still touch the dashboard for meaningful parts of the workflow.

Who Should Use Loops Now — and the Alternatives

Loops in 2026 is a strong fit for a specific team: a SaaS company where engineers own the email integration, the product emits lifecycle events, and someone on the team already builds with AI agents. For that team, the CLI, Skills, and Goals releases compound into a genuinely faster setup. For everyone else, the fit is weaker.

If you need heavier lifecycle orchestration and behavioral segmentation, Customer.io is the closest competitor and a common upgrade path once a company outgrows Loops. It is far more powerful, but also more complex to set up and more expensive, with a steeper learning curve that can stall a small team.

If your need is purely transactional — password resets, receipts, notifications with no marketing layer — a specialist like Postmark delivers faster support and a longer deliverability track record, though it deliberately does not do marketing campaigns at all, so you would still need a second tool for newsletters. For a wider look at the category, our roundup of the best email marketing tools for SaaS covers where Loops sits against the full field.

The pattern worth noting is bigger than Loops. Kit and Omnisend both shipped AI integrations in spring 2026, and the entire category is racing to make itself drivable by the assistant a team already uses — a shift covered in our piece on Omnisend’s MCP launch. Loops took the developer-tools route instead of a chat connector: a CLI, open-source tooling, and Skills that live where code is written. Same destination, different road. For a product-led SaaS team, Loops’ road is often the shorter one.

Sources

1. Loops — Pricing — accessed 2026-07-21
2. Loops — CLI (GitHub) — accessed 2026-07-21
3. Loops — Agent Skills — accessed 2026-07-21