Workflow Automation

n8n204.7k★

N8n is the open source alternative to Zapier. It gives you a visual workflow builder where you drag nodes, connect them, and watch data flow through. Over 400 integrations built in. Self-hosting is free under a 'sustainable use' license (not fully open source, you can't resell it as a service). The cloud version starts at $24/mo. The self-hosted version has no feature restrictions, no execution limits, and no user limits. The catch: setting up n8n in Docker and keeping it running takes real skill. Updates can break workflows. The 'sustainable use' license means you can't build a competing automation platform on top of it. And while the visual builder is powerful, complex logic (loops, error handling, conditional branches) gets messy in a node graph. Sometimes a Python script is cleaner than 40 connected nodes.

activepieces24.6k★

Activepieces is Zapier you can run yourself. Build automations by connecting triggers to actions across 280-plus integrations, with loops, branches, retries, and a code step that takes real npm packages when the visual builder runs out. The interesting turn is that every one of those integrations is also exposed as an MCP server, so the same catalog your no-code flows use can be handed to Claude, Cursor, or any MCP client. Integrations are called pieces, they are TypeScript npm packages, and roughly 60% of them come from the community. That matters for the usual open-core complaint: when a connector is missing you write it yourself with hot reloading, instead of filing a request and waiting. Self-hosting is Docker plus Postgres, and it is network-gapped by design, which is the actual reason regulated teams pick it over Zapier. The core is MIT and free with no seat cap when you host it. Cloud starts free for one person, then $16/mo for up to five users and $166/mo for a team of 25. The catch is the license split. Everything under packages/ee carries a separate commercial license, so the enterprise governance and branding pieces are not MIT. Read that directory before you build a product on top of it.

Windmill18.0k★

Windmill turns your scripts, in Python, TypeScript, Go, Bash, or SQL, into scheduled jobs, API endpoints, workflows, and internal tools with real UIs. Write a function and Windmill generates the input form, then handles scheduling, retries, permissions, and logging. It is what you get when n8n meets a proper IDE. The self-hosted Community edition is free under AGPLv3, and the limits are more generous than the marketing suggests: up to 50 users, with SSO capped at 10 of them. Unlimited executions, 3 workspaces, unlimited scripts, flows and apps, and a maximum of 4 permission groups. Self-hosting runs on Docker Compose with Postgres and takes under an hour. A 2GB RAM VM handles it comfortably, and the ongoing ops burden is mostly database maintenance. Enterprise starts at $120 a month, and how it prices matters more than the headline. Seats are $20 a month per developer and $10 per operator, where an operator can run scripts and flows but not create or edit them. Compute is billed separately at $50 per worker per month, and a group of 8 native subworkers is another $50. Their own calculator quotes $170 a month for a single developer with three compute units, so treat $120 as a floor rather than a plan price. Solo devs: free covers everything. Small teams: free, and 50 users goes further than most open core limits. Teams that need SSO past 10 people, audit logs, or SAML group sync: that is the Enterprise line. The catch: Windmill is less known than n8n or Retool, so the community is smaller and there are fewer pre-built integrations and examples to copy. If you need hundreds of ready-made SaaS connectors, n8n is ahead. If you would rather write actual code than drag boxes around, Windmill wins that trade easily.

Argo Workflows17.0k★

Argo Workflows orchestrates multi-step jobs on Kubernetes: data pipelines, CI/CD, ML training, batch processing, all defined as directed acyclic graphs in YAML. Each step in your workflow runs as a container on your cluster. You get parallelism, retries, conditionals, loops, and artifact passing between steps. The web UI shows you the DAG visually with real-time status. It's a CNCF graduated project, which means serious production adoption. Used by Intuit, Tesla, Google, and many others. Completely free. Apache 2.0 license. Akuity offers commercial support and Argo CD (related but separate) as a managed service, but Argo Workflows itself is free. Small teams already on Kubernetes who need workflow orchestration beyond simple CronJobs should look here. If you're NOT on Kubernetes, this isn't for you. The entire model assumes K8s. The catch: you need Kubernetes. That's a hard prerequisite. And YAML workflow definitions get painful fast for complex pipelines. The learning curve combines Kubernetes complexity with workflow engine complexity. Not for the faint of heart.

camel6.3k★

Camel has been wiring enterprise systems together since 2007, and the reason it survived is 350-odd connectors you do not have to write yourself. It is an integration framework: it moves data between systems that were never designed to talk. A route reads from one place (a database, a Kafka topic, an FTP drop, an HTTP endpoint) and delivers somewhere else, with transformation and error handling in between. Routes are written in Java, YAML, or XML. Apache 2.0, free. Deployment is flexible: Spring Boot, Quarkus, or standalone through the Camel CLI. Java is the hard requirement. Kaoto and Karavan give you visual route design if hand-writing the DSL is not your thing, and there is an MCP server now, so an AI assistant can work with your routes directly. Free, entirely, with no paid tier from the project. Solo developers rarely need this. It earns its place once you have more than a handful of systems exchanging data and you are tired of bespoke glue scripts that fail silently at 3am. The catch is weight and learning curve. Enterprise Integration Patterns are a real body of knowledge and Camel assumes you know them. For two or three simple integrations, a few scripts or an n8n workflow gets you there far faster. Camel pays off at scale and charges you upfront in understanding.

camunda4.3k★

Camunda orchestrates business processes across people, systems, and devices. BPMN workflow engine, human task management, monitoring dashboards, and analytics, all in one platform. Organizations running complex multi-step processes that involve both automated systems and human decisions will find this built for exactly that problem. The core engine (Zeebe) is a high-throughput, horizontally scalable workflow engine built on gRPC. Self-hosting on Kubernetes is the standard deployment path. It handles millions of workflow instances and integrates with Kafka and standard message queues. The operational tooling (Operate, Tasklist, Optimize) gives you visibility into running processes and human task queues. Enterprise teams with complex approval workflows, order processing, or compliance processes get the most value. Solo developers and small teams will find it overkill unless they're specifically building process orchestration. The cloud offering removes the ops burden but comes at enterprise pricing. The catch: the community edition is real and usable, but the best operational tooling is gated behind enterprise licensing. And running Zeebe on Kubernetes is not trivial ops work.

dagu4.1k★

dagu runs scheduled jobs and multi-step workflows from a single binary, with no database or message broker required. You define what runs in plain YAML: shell scripts, Docker containers, Kubernetes jobs, SSH commands, HTTP calls, and coding-agent CLIs like Claude Code or Codex as steps. It ships a web UI for watching runs and digging into logs, and the core is free under GPLv3. It is cron that grew up, with a dashboard and retry logic. Self-hosting is about as light as orchestration gets. Drop the binary on a Linux, macOS or Windows server, point it at a directory of YAML files, and you have a scheduler. State lives in files, so there is no Postgres or Redis to babysit. Workers can be long-lived machines or temporary Cloud Run and Kubernetes jobs, and a single instance handles real work for most teams. Solo developers and small teams replacing a pile of crontab entries will love this, and the free Community edition covers the job running. Its limits are team features: two API keys, no user management, no SSO, RBAC or audit logs. A self-host license adds those at $50/mo for three servers or $150/mo for fifteen, workers unlimited. There is no hosted version anymore. That is the fair kind of paywall. The catch is that YAML workflows get unwieldy fast. Simple pipelines stay readable, but complex branching logic starts to fight you. If your orchestration needs are heavy, look at Temporal or Airflow before you outgrow this.

viewflow2.9k★

Viewflow turns Django into a platform for business apps with real process automation built in. You define BPMN-style workflows in Python (approvals, multi-step processes, who does what next) and get the surrounding pieces too: CRUD screens, user management, reporting. If you already build on Django, it saves you from hand-rolling a state machine and an admin UI for every internal process. The core is open source. The catch is the licensing, two layers of it. Viewflow Core is AGPL-3.0, which is fine for internal tools but worth a lawyer's glance before you ship it inside a closed-source product. The deeper integrations and the right to keep private forks live in Viewflow PRO, the paid commercial version. So free gets you a working BPMN engine; the convenience and the permissive license cost money. Solo developers and small teams building internal workflows: the free Core is plenty, run it yourself. Teams shipping a commercial product on top of it, or who want the prebuilt integrations and AGPL out of the way, should treat PRO as the honest path, and it is priced as a developer tool, not enterprise extortion. If you are not already on Django, this is not a reason to switch; pick a workflow engine that matches your stack instead. One more thing: this is a framework, not a no-code app. Your team needs Django chops to get value out of it. Hand it to non-developers expecting a Zapier-style canvas and they will bounce.