Open-source Klaviyo alternatives
There is no open-source Klaviyo. Nothing self-hostable models orders, products, browsing sessions and revenue-per-flow the way an ecommerce platform does, and pretending otherwise would waste your afternoon.
What does exist is open-source email that a store can use for the simpler jobs. Below is what each project actually gives you, and where the gap starts.
What each one asks of you
Read this table knowing that none of these has an ecommerce data model. The question is how much of your programme needs one.
| Provider | Licence and cost | What it asks of you | Type |
|---|---|---|---|
| Mautic | Open source | PHP, MySQL and reliable cron. Upgrades have a reputation. You bring the sending provider | Self-hosted |
| listmonk | AGPL-3.0 | One Go binary plus Postgres — the easiest here to run. No automation builder | Self-hosted |
| Keila | AGPL-3.0 | Elixir and Postgres, or a European hosted plan from €8/mo. No lifecycle automation | Self-hosted |
| Ghost | Open source (MIT) | Node and a database. Publishing and memberships rather than marketing automation | Self-hosted |
| Plunk | AGPL-3.0 | Postgres, Redis and a container, or $0.001 an email on Plunk Cloud | Self-hosted |
| useSend (formerly Unsend) | AGPL-3.0 | Postgres, Redis, your own AWS account — and it's still described as beta | Self-hosted |
| Notifuse | AGPL-3.0 | Postgres and a Go binary, plus a sending provider underneath. Cloud from $16/mo | Self-hosted |
| FluentCRM | $90/year, one site | Lives inside WordPress. Site down means marketing down; large sends strain WP cron | Self-hosted |
| Sendy | $69 one-time | PHP and MySQL on your own hosting, sending through your own Amazon SES account | Self-hosted |
| Mailcoach | Separate self-host licence | Hosted is $9.99/mo plus per-email; the self-hosted licence isn't published | Self-hosted |
| Postal | MIT | A full mail server — you own the IPs, the warm-up and the blocklist relationships | Self-hosted |
| Email Bump | Proprietary | Not open source and not self-hostable. The trade is that nobody pages you about a blocklist | Permanent |
From each project's own README, documentation and pricing page. Licences change; check before you build on one.
Nothing open source knows what an order is
Klaviyo's value is a data model: every order, product, line item and browsing session attached to a profile, with revenue attributed back to the flow that produced it. Abandoned cart works because it knows what was in the cart. Back-in-stock works because it knows the inventory.
No open-source project on this page has that. Mautic comes closest through custom fields and integrations, and you would be building the ecommerce model yourself — mapping order webhooks onto contact fields, writing the segment logic, and forgoing revenue attribution entirely.
For a store doing real revenue, that's a poor trade. The genuine cost saving is Klaviyo's per-profile pricing, and the genuine cost is rebuilding a data model someone else already maintains.
Self-hosting suits the simple half
Plenty of stores run one recovery sequence, a post-purchase thank-you and a monthly newsletter. All of that is expressible with an event and a segment, and Mautic or Plunk will do it on your own infrastructure with no per-contact cost at any list size.
The test is the same one on the main page: open your flow reports and count how many automations branch on product or order data rather than on 'did this event happen'. If the answer is one or two, self-hosting is viable. If your merchandising lives inside the email, it isn't.
The middle path most stores land on is Omnisend — an actual ecommerce data model at a fraction of Klaviyo's price — rather than self-hosting at all.
If you do it anyway, know what you’re taking on
You'd be building the data model. Mapping order webhooks onto contact fields, writing the segment logic by hand, and doing without revenue attribution entirely. That's real engineering work maintaining something Klaviyo already maintains for you.
And owning deliverability. A store's transactional and marketing mail both matter, and self-hosting means blocklist removals and warm-up schedules land on whoever is nearest. For a business where email drives revenue, that's a considerable risk to absorb for a licence saving.
The licences split two ways. Mautic, listmonk, Keila, Plunk, useSend and Notifuse are copyleft; Ghost, Postal and FreeResend are permissive. Read properly before embedding any of them in something you sell.
Klaviyo alternatives, asked and answered
Is there an open-source Klaviyo alternative?
Not really, and we'd rather say so. No self-hostable project models orders, products and revenue attribution the way Klaviyo does. Mautic is the closest general-purpose option and you'd be building the ecommerce layer yourself.
If price is the driver rather than ownership, Omnisend is the realistic answer — the same shape as Klaviyo, materially cheaper.
Can Mautic do abandoned cart?
It can fire a sequence when you send it an event, which covers a basic 'you left something behind' email. What it can't do without significant custom work is put the cart contents in the message or attribute the recovered revenue.
That's the difference between an abandoned-cart email and an abandoned-cart programme.
What if I just want to stop paying per profile?
Then the question is whether your profile count is inflated. Klaviyo bills on active profiles, which is broader than people who opted into marketing — stores syncing every customer record pay for guests who never subscribed.
Auditing that often saves more than a migration would, and costs an afternoon rather than a rebuild.
Behavioural email without the ecommerce tax
Track your own events, build live segments from them and run multi-step flows — priced on contacts, with sending uncapped. No ecommerce data model, and we say so.