Fast self-hostable open-source workflow engine

Original Article ↗ Hacker News Discussion ↗

A new self-hostable workflow engine called Windmill is drawing attention for combining fast, Rust-based execution with a low-code web UI, script editor, and app builder aimed at automating internal business processes. Commenters praise its developer experience, PostgreSQL-backed queueing, and flexibility for tasks ranging from ETL and data pipelines to cron replacement and internal tools, while comparing it to Airflow, Dagster, Temporal, Retool, and Rundeck. There is active debate around its positioning (workflow engine vs. internal app platform), licensing model (AGPL with proprietary enterprise plugins), and how well it fits enterprise environments already invested in other orchestration and automation stacks.

Overall Reception

  • Many commenters are very positive, calling Windmill one of their favorite self‑hosted tools and praising its speed, reliability, and active support.
  • Others find the marketing and positioning confusing and are unsure what problems it is meant to solve relative to other workflow / internal tools products.

Use Cases & Strengths

  • Reported real‑world uses: daily content generation, PDF and video automation, internal tools, web crawlers, yt‑dlp jobs, cron replacement, simple ELT/data workflows.
  • Strong fit for internal, developer‑centric tooling: background jobs, APIs, dashboards, approvals, event‑triggered scripts.
  • Several people emphasize that the auto‑generated UIs, logging/observability, and web UI for scripts solve the pain of ad hoc cron + shell + scattered monitoring.

Performance & Architecture

  • Discussion around whether “fastest workflow engine” matters: consensus that sub‑second latency improves developer experience and enables event‑streaming / low‑latency use cases, but isn’t for HFT.
  • Backend uses PostgreSQL as the job queue; this sparks debate:
    • Pro: simplifies backups, good enough for typical job tables, aligns with modern Postgres‑queue libraries.
    • Con: concerns about connection limits, locking, scaling, and “NIH vs using dedicated queues.”
  • Benchmarks and graphs are questioned for missing axis units; clarified as seconds with more detail in docs.

Positioning vs Other Tools

  • Compared and contrasted with: Airflow, Dagster, Prefect, Temporal, Rundeck, Pipedream, Retool, Budibase, node‑red, n8n, BPMN engines, even ERP/RPA systems.
  • Some see it as:
    • Primarily a workflow engine with strong scripting and a developer‑oriented app builder.
    • A potential replacement for Airflow/Rundeck in simpler pipelines and scheduled jobs.
  • Others argue it’s not ideal for complex or highly dynamic data/ETL pipelines (where Dagster/Airflow still win), though Windmill’s team claims recent features (S3, restartable flows, ETL ergonomics) improve this.

Fit for Business Processes

  • For complex, human‑heavy business workflows (emails, calendars, approvals, customer communication), opinions diverge:
    • Some think Windmill can work well if you’re willing to code integrations and gradually automate steps.
    • Others recommend domain‑specific “pipeline” stacks (e.g., VFX‑style tools, CRMs, BPMN engines) for richer human‑workflow and calendar‑driven use cases.

Licensing, “Open Source” Status & Enterprise Adoption

  • Core is AGPL; SSO and other “enterprise plugins” are source‑available and free up to 10 users.
  • Clarifications:
    • Triggering workflows via API does not force your app to be AGPL; white‑labeling/extending Windmill itself does.
    • Self‑hosted enterprise version is licensed by vCPU but does not phone home (usage reporting is discussed, not yet active).
  • Some see this as open‑core bordering on “source‑available”; others are comfortable with AGPL + paid enterprise.

Developer Experience & Usability Feedback

  • Praised: online code editor with LSP, VS Code extension, git sync flows, YAML/OpenAPI “OpenFlow” spec, support for Python, TypeScript, Go, Bash, SQL; upcoming Ruby; possible Rust.
  • Mixed: spell‑checking and minor typos in marketing content; unclear pricing tiers and feature gating (e.g., content search); no localization yet.
  • Several suggest clearer messaging and copywriting to sharpen the value proposition and target audience.