Launch HN: Creo (YC W24) – Build internal tools with NextJS and AI

Hacker News Discussion ↗

Product overview & workflow

  • Tool positions itself as a way to build internal tools (dashboards, CRUD apps) using an opinionated Next.js starter plus an AI assistant.
  • Typical flow: run a CLI to init a Next.js boilerplate with their UI package, develop locally with an AI chat pane that edits files, push to GitHub, then connect the repo in their hosted app to get auth, orgs, and permissions “wrapped around” the tools.
  • Users connect their own databases; the platform does not currently offer managed DBs.

Value proposition & differentiation

  • Claimed value:
    • Avoiding setup for auth, RBAC, org/team management, and UI consistency.
    • Opinionated stack (Next 14, React Server Components, shadcn-based UI, their own CSS/layout rules) that the AI “knows” deeply.
    • Fast feedback: natural-language changes to tools instead of manual UI wiring.
  • Many commenters say this looks like “just” a Next.js boilerplate + common auth + open-source components, with unclear differentiation from simply using Next on Vercel plus an auth provider.

Pricing & business model

  • Pricing around $30/month per user is widely criticized as misaligned with developers who can assemble similar stacks themselves.
  • Multiple suggestions: open-source the boilerplate and charge only for the AI layer.
  • Some argue that if it meaningfully saves a developer month of work, subscription could be rational, but others note ramp-up, lock-in, and ongoing cost.

AI assistant debate

  • Supporters:
    • Component-aware AI with enforced design rules can outperform generic GPT prompts that often hallucinate shadcn APIs or misuse Next 14 patterns.
    • Tight integration (can write to disk, react to compile errors) aims to streamline the common “paste into GPT, paste back” workflow.
  • Skeptics:
    • IDE agents (Copilot, Cursor, Continue) already provide similar workflows.
    • AI-generated UIs often fail on real-world edge cases; prompt-tweaking may erase any time savings.
    • Some find coding in a browser UI less ergonomic than staying in their editor.

Tech stack choices & internal-tools fit

  • Proponents say Next.js + RSC are “LLM friendly” due to simpler APIs and clear server/client boundaries.
  • Critics question using a web-focused framework for internal tools that rarely need SEO and worry about added complexity and vendor lock-in.

Alternatives & competition

  • Commenters repeatedly compare it to Retool, Django admin, Supabase/Clerk/Auth0 + boilerplates, Refine, PocketBase, Streamlit, and open-source tool builders (e.g., ToolJet).
  • Many feel the non-AI parts are already “solved” and commoditized, making differentiation and pricing an uphill battle.

Security, deployment, and data access

  • Some confusion remains about whether the hosted environment gets direct access to customer databases; it’s clear users must bring their own DB, but the exact connection/security model is seen as underexplained.

Perceptions of market fit & YC backing

  • Several see the space as crowded with mature free/OSS options and question the startup’s defensibility.
  • Others note YC often backs teams pre-pivot and that strong founders plus some traction may have driven the funding.
  • A few users express genuine interest, especially those who like opinionated stacks and dislike low-code builders; others see it as a “nice toy” rather than something they’d risk for production.

Naming & trademark concerns

  • Multiple comments flag potential confusion or trademark conflicts with existing products named “Creo” in CAD/printing software; founders acknowledge they will look into it.