Skip to content
Documentation

AI Philosophy

We’re unashamedly pro-AI at Norri — but not in the “AI slop” sense. We believe AI is a genuinely useful tool when applied thoughtfully, and we use it extensively in ways that benefit both our development process and your experience.

AI in Development

We use AI as an important development tool:

  • Code assistance — AI helps us write, review, and improve code faster
  • Documentation — Much of this documentation was drafted with AI assistance
  • Problem solving — AI helps us explore solutions and catch issues early
  • Both local and cloud — We use a mix of local models and cloud AI services during development

We’re transparent about this because we think it’s the right way to build software in 2024 and beyond. AI doesn’t replace thoughtful engineering — it amplifies it.

AI in Norri’s Features

When it comes to AI features within Norri itself, our approach is fundamentally different:

Privacy First

Any AI-powered features in Norri will:

  • Never use cloud AI providers — No data leaves your server
  • Never connect externally — No API calls to OpenAI, Anthropic, or anyone else
  • Never share your data — Your library, watch history, and preferences stay on your hardware
  • Use embedded local models — If AI is needed, it runs entirely on your server

Planned AI Features

We’re exploring features like:

  • Semantic search — Describe what you’re in the mood for in natural language (“something like Blade Runner but more upbeat”)
  • Smart recommendations — Suggestions based on your library and viewing patterns using local embeddings
  • Content analysis — Automatic tagging and categorization

All of these would run locally using embedded models — no cloud, no data sharing, no privacy concerns.

Why This Matters

Many “AI-powered” products are just thin wrappers around cloud APIs. Your data gets sent to third parties, processed on their servers, and potentially used to train their models.

Norri is different. We believe you shouldn’t have to choose between smart features and privacy. By running AI locally:

  • Your data never leaves your network
  • Features work offline
  • No subscription fees for AI capabilities
  • No dependence on external services that could change or disappear

Our Commitment

We will never:

  • Send your library data to external services
  • Require cloud connectivity for AI features
  • Use your data to train external models
  • Add AI features that compromise your privacy

AI should work for you, on your terms, on your hardware.