From empty folder to production.

Spin up a full local stack in minutes

A fully dockerized full-stack environment — Next.js frontend, NestJS/Prisma backend, Postgres — up and running in minutes and ready to build on. Run the whole system end-to-end on your machine: frontend, API, database, secrets, the works. No half-mocked services, no cloud round-trips to test a change.

It also ships with a CLAUDE.md — built-in instructions that teach your agent how to build on the scaffold: test-first, clean code, clean architecture, the same best practices baked into the stack itself. Claude stays on the rails from the very first commit instead of improvising its own conventions.

Fully dockerized, fully portable

Everything is containerized from day one, so the exact same image runs on your laptop and in production. No lock-in, no rewrite to deploy elsewhere — it runs anywhere Docker runs, on any cloud.

Real-time, not just REST

Plain request/response only gets you so far. squidpush wires in real-time messaging out of the box — live updates, presence, and broadcast channels over the Pusher protocol, hosted (Pusher.com) or self-hosted with Soketi in your own infra. Build live dashboards, chat, collaborative tools, push notifications, and progressive mobile apps — the interactive experiences users actually expect — instead of stopping at a basic CRUD API.

Durable workflows & background jobs

Real apps do work outside the request — scheduled tasks, retries, multi-step flows that survive a crash or a deploy. squidflow sets up durable workflows and background jobs with Inngest, running locally in Docker for dev and on the cloud version in production. Long-running jobs, fan-out, retries, and step functions — without hand-rolling a queue or a cron server.

Push to main → it ships

The CI/CD skill builds you a pro-grade pipeline that builds, tests, and deploys to production automatically on every push to main. Pick your provider — it generates the right pipeline for each:

It doesn't just run — it scales

Production isn't about surviving traffic — it's about thriving under it. The deploy skills bake in performance and scaling best practices: aggressive caching at every layer, horizontal scaling that adds capacity as load climbs, and right-sized resources throughout — so your app runs as fast as it possibly can while costing as little as it possibly can. Scale up when the traffic is there; scale down and stop paying when it is not.

Going live isn't the finish line

A deployed, working app in the cloud is the start, not the end. The observability skill wires in full monitoring — dashboards, health checks, and alerts — so you know the moment something drifts, before your users do. Production you can actually see.

Your agent does the heavy lifting

With the ops skill, your agent configures, sets up, and accesses every tool in the stack — cloud CLIs, secrets, the database, the registry — and cross-references them so they actually work together. You describe the outcome; it wires up the plumbing.

Strictly opinionated — by design

Every skill encodes best practices: clean code, clean architecture, sensible defaults, and the guardrails a staff engineer would insist on. It deliberately goes far beyond the bare requirement — because the gap between "works" and "production-grade" is exactly what you're buying.