Public roadmap for the StreamMind AI project. Updated regularly.
Foundation release. Extension-first, single LLM provider, hallucination prevention.
- Monorepo architecture (pnpm + Turborepo)
- Chrome Extension (Manifest V3)
- React + TypeScript + Vite + Tailwind
- BYOK flow (Bring Your Own Key)
- OpenAI adapter implementation
- TMDB catalog integration
- LLM Orchestrator with retry logic
- Strict Zod schema validation
- Hallucination detection and prevention
- Input sanitization (prompt injection guard)
- Dark mode elegant UI
- Skeleton loaders + error states
- Movie cards with AI reasoning
- Fastify backend with rate limiting
- CORS + Helmet security
- Production-grade folder structure
- README, ARCHITECTURE.md, LLM_DESIGN.md
Expand LLM support, add caching layer, improve quality.
- Anthropic Claude adapter
- Google Gemini adapter
- Provider selector in extension UI
- Redis caching layer for TMDB data
- Response caching (same query → cached result)
- Improved catalog sourcing (trending, top-rated, genre-based)
- Better recommendation ranking algorithm
- Provider quality comparison metrics
- Unit tests for all packages
- Integration tests for API routes
- CI/CD pipeline (GitHub Actions)
- Docker Compose for local development
Polish the user experience, add personalization hooks.
- Streaming responses (real-time token rendering)
- Search history in extension
- Favorite movies list
- "Not interested" feedback mechanism
- Genre-based quick filters
- Multi-language support (i18n)
- Firefox extension support
- Extension settings page (advanced config)
- Accessibility audit (WCAG 2.1 AA)
- Performance monitoring (Sentry/Datadog)
Richer catalog data, better context for LLM.
- Movie streaming availability (via JustWatch or similar)
- Watch provider logos on movie cards
- Director/cast information in prompts
- Movie trailers integration
- Similar movies API endpoint
- Region-specific catalog filtering
- TMDB image CDN optimization
- Catalog refresh scheduler
Production-grade, scalable, ready for commercialization.
- PostgreSQL for user data
- Redis for session + rate limiting
- Vector database (Pinecone/Weaviate) for semantic search
- API Gateway (Kong/AWS API Gateway)
- Multi-region deployment
- CDN for static assets
- User accounts and authentication
- Taste profile learning (collaborative + content-based)
- Hybrid recommendation engine (LLM + traditional ML)
- Embeddings-based movie similarity
- Watchlist with notifications
- Social features (share recommendations)
- Premium tier (higher rate limits, faster models)
- Billing integration (Stripe)
- Usage analytics dashboard
- A/B testing framework
- Affiliate links to streaming platforms
- Partner API for third-party integrations
- TV Shows — Expand beyond movies to series recommendations
- Mobile App — React Native companion app
- Browser-Native AI — Leverage Chrome's built-in AI APIs
- Community Curations — User-generated recommendation lists
- AI Movie Reviews — Personalized review summaries
- Global Expansion — Support for 20+ languages
This roadmap is subject to change based on community feedback and contributions. Want to contribute? Check out our Contributing Guide.