Zapier bills per task, stores your OAuth tokens on their cloud, and quietly becomes line item #47 on the credit card nobody owns. You do not need a finance committee to connect "new form row" to "Slack ping" — you need a workflow tool that runs on hardware you control and does not charge per zap.
Automatisch (~14k stars) is an open-source Zapier-style automation platform: visual flows, triggers and actions across apps (Slack, Twitter/X, email, webhooks, and a growing integration list), no code required for the happy path. The community edition is AGPL-3.0; the same repo also ships enterprise-licensed files (anything with .ee. in the filename) if you need commercial features later.
What you actually get
Automatisch is built for operators who think in "when this happens, do that" — not in JSON graphs and custom nodes unless they want to.
- Flow builder — connect services with a UI similar to mainstream iPaaS tools
- Integrations — Twitter, Slack, and more; check their docs for the current app catalog
- Webhooks — receive events from your own apps and fan out to other steps
- Self-contained stack — web app plus background worker, Postgres for state, Redis for queues
We already covered n8n on this blog. Rough split: n8n is the power-user Swiss Army knife (hundreds of nodes, code steps, AI hooks). Automatisch is closer to "we used Zapier until the bill and the data residency questions got awkward." Many teams run one or the other, not both, unless you are deliberately separating marketing automations from engineering pipelines.
Why self-host?
Credentials stay home. Every OAuth connection for Slack, mail, or social sits in your database on your VPS — not in a US SaaS subprocessors spreadsheet.
GDPR and sector rules. Healthcare, finance, and EU-facing shops often cannot ship customer events to arbitrary clouds. Canadian hosting on a VPS you name is an easier story than "trust Zapier's DPA."
Predictable cost. You pay for the server and your time, not per-execution metering that spikes when marketing turns on a new funnel.
No lock-in. AGPL CE means you can export flows and data and move — harder than escaping a proprietary zap history, easier than rewriting everything in bash.
Running it
Official path: clone the repo and docker compose up. That brings up:
- main — UI on port 3000
- worker — executes flows (same image,
WORKER=true) - Postgres 14 and Redis 7
Set real secrets before production: ENCRYPTION_KEY, WEBHOOK_SECRET_KEY, APP_SECRET_KEY, plus Postgres passwords. Default demo login is user@automatisch.io / sample — change that immediately in settings.
Put TLS and a hostname in front (Caddy, nginx, Traefik); set HOST, PROTOCOL=https, and PORT to match your proxy. Plan 2–4 GB RAM for a small team instance; workers and Postgres grow with concurrent flows and integration count.
Back up Postgres and the automatisch_storage volume — flow definitions and uploaded assets live there.
Who it's for
Good fit: small business ops, agencies wiring lead forms to chat, founders who outgrew manual copy-paste but do not want to learn n8n yet.
Maybe skip: heavy ETL, sub-second event streams, or teams that need enterprise-only Automatisch EE features without reading the license split.
We host the box — you bring domains and integration accounts. Docker hosting in Canada with snapshots suits the compose stack. Tell us how many flows run concurrently and which SaaS APIs you connect and we will size Postgres and workers before you paste production OAuth secrets.
