Windmill.dev

Operator's take

The setup most operators land in is this: someone on the team has scripts — Python, TypeScript, Go, Bash, SQL, whatever — that work fine locally but are completely inaccessible to everyone else. Turning a script into something the ops team can actually trigger means building a UI, standing up an endpoint, writing error handling, adding logging. That's weeks of plumbing work that never touches the original business logic. Windmill's bet is that all of that plumbing should be automatic. You write the function signature; Windmill generates the form, the API route, the scheduler, and the execution logs around it.

Where it earns its keep is the gap between "we have scripts" and "we have tools." If your team is running manual data syncs, trigger-on-demand reports, or ETL jobs that someone has to SSH in to run, Windmill turns those into self-service dashboards that non-technical staff can operate without ever touching a terminal. The visual workflow composer also lets you chain scripts into multi-step pipelines — conditional branches, retry logic, dependency ordering — without writing orchestration code on top. Windmill has also added first-class AI agent support: you can build and orchestrate AI agents as flows alongside your existing automation, using the same worker infrastructure and observability tooling.

The honest constraint: this is not no-code. If nobody on your team writes code, Windmill won't help you much — the auto-generated UI only works as well as the code going in. The generated forms are functional and clean but not custom-designed for casual use; the app builder lets engineers construct polished interfaces with drag-and-drop components, but pixel-perfect client-facing apps are still more work than a dedicated frontend. And if your automation needs are mostly connecting SaaS apps to each other rather than running code, n8n or Make will get you there faster with less setup. Windmill's sweet spot is developer or data-engineering teams who want to expose their existing scripts as reliable, observable internal tooling without building a custom backend around each one.

What it's good at

• Auto-generated UI from code — function parameters become form fields automatically; no frontend development needed to make a script accessible to non-technical users.
• Instant API endpoints — any script becomes a REST endpoint with built-in auth, rate limiting, and error handling in a single step.
• Cron and trigger scheduling — schedule jobs at intervals, watch for triggers, and monitor every run through a centralized dashboard with full execution logs.
• Visual workflow composer — chain scripts into multi-step pipelines using a drag-and-drop editor or YAML; supports conditional branches and inter-step dependencies without orchestration code.
• AI agent orchestration — build and run AI agent workflows on the same infrastructure as your other scripts and flows; first-class use case on the current platform.
• Multi-language support — Python, TypeScript, Go, PHP, Bash, SQL (PostgreSQL, BigQuery, Snowflake, MS SQL on Enterprise), and 20+ additional language runtimes.
• Self-hostable with full observability — runs on your own infrastructure for data compliance; enterprise-grade reliability and logs baked in.
• Free and open-source core — self-hosted Community Edition is free and open-source (up to 50 users, 3 workspaces); Enterprise tier starts from $120/mo (cloud) with audit logs, SAML, S3 cache, and priority support.

What it's not

• Not a no-code tool — requires coding knowledge (Python, TypeScript, Go, Bash, SQL, or similar); if no one on your team writes code, there's nothing for Windmill to wrap.
• Not a turnkey UI builder — the auto-generated forms are functional but minimal; the drag-and-drop app builder gives engineers more control, but a polished client-facing interface still needs real frontend investment.
• Not a SaaS integration platform — if your workflow is mostly connecting Salesforce to Slack to Google Sheets, n8n or Make will get you there with less friction than Windmill.
• Not a data warehouse or BI layer — Windmill orchestrates the pipeline; it doesn't store or visualize the data at the end of it.