Product Roadmap
This page provides a complete view of where Carbon Connect has been, where it is now, and where it is going. The platform has been built through seven structured development phases, with clear priorities for the near, medium, and long term.
Completed Development Phases
Phase 1: API Layer Foundation
The core platform infrastructure: user authentication, multi-tenant architecture, and company management.
| Deliverable | Status |
| FastAPI application with 60+ endpoints | Complete |
| JWT authentication with access and refresh tokens | Complete |
| Multi-tenant middleware with row-level security | Complete |
| Company CRUD operations with pagination and filtering | Complete |
| OAuth2 password flow | Complete |
| Token blacklisting for secure logout | Complete |
Phase 2: Grant Search and Discovery
Full-text search, filtering, and grant management capabilities.
| Deliverable | Status |
| Meilisearch full-text search (sub-100ms) | Complete |
| PostgreSQL fallback search | Complete |
| Filtering by country, NACE codes, funding range, status, size, source | Complete |
| Grant statistics and analytics endpoints | Complete |
| Pagination with configurable page size | Complete |
| Deadline tracking with upcoming deadlines view | Complete |
Phase 3: Matching Engine
The hybrid scoring algorithm that powers intelligent grant recommendations.
| Deliverable | Status |
| Five-component hybrid scoring algorithm | Complete |
| 30% rule-based criteria scoring | Complete |
| 25% semantic similarity via embeddings | Complete |
| 25% carbon alignment scoring | Complete |
| 10% collaborative filtering | Complete |
| 10% recency/urgency scoring | Complete |
| Match persistence and refresh | Complete |
| User interaction tracking (view, save, dismiss) | Complete |
Phase 4: Carbon Intelligence
Carbon profiling, emission calculations, and carbon-enhanced matching.
| Deliverable | Status |
| Anthropic Claude API integration (Sonnet 4, Opus 4, Haiku 3.5) | Complete |
| Climatiq API integration for emission factor calculations | Complete |
| Carbon-enhanced matching (25% carbon weight) | Complete |
| Carbon category classification (14 categories) | Complete |
| EU Taxonomy alignment scoring | Complete |
| Certification tracking (ISO 14001, SBTi, CDP, B Corp) | Complete |
| GHG Protocol Scope ½/3 support | Complete |
| Carbon database schema (companies + grants) | Complete |
Phase 5: Application Assistant
AI-powered grant application drafting using Claude.
| Deliverable | Status |
| Claude-powered application generation | Complete |
| Section templates for common grant applications | Complete |
| Content filter handling and safety checks | Complete |
| Token usage tracking for cost monitoring | Complete |
| Streaming support for real-time generation | Complete |
| Cost estimation per model tier | Complete |
Phase 6: Frontend Dashboard
Complete user interface for all platform features.
| Deliverable | Status |
| Next.js 14 with App Router | Complete |
| 40+ React components | Complete |
| Dashboard, Grants, Companies, Applications views | Complete |
| Carbon profile forms and visualizations | Complete |
| Tailwind CSS + shadcn/ui design system | Complete |
| React Query for server state management | Complete |
Phase 7: Infrastructure and Operations
Background processing, email notifications, storage, and cloud infrastructure.
| Deliverable | Status |
| Celery background workers with 4 queues | Complete |
| Amazon SES email integration with templates | Complete |
| S3 document storage with lifecycle policies | Complete |
| Celery Beat scheduled sync tasks | Complete |
| Terraform infrastructure modules (VPC, RDS, ElastiCache, S3, ALB, ECS) | Complete |
| AWS Secrets Manager integration | Complete |
| Meilisearch service with index management | Complete |
| Notification system (database + service + API) | Complete |
| Partner API with 14 endpoints | Complete |
| Reference data API (NACE codes, countries) | Complete |
| Dashboard API with tenant-scoped aggregations | Complete |
| Grant pipeline Celery tasks (embedding + deduplication) | Complete |
| Docker deployment configuration | Complete |
Current Status: MVP Finalization
The platform is feature-complete for MVP. The remaining work focuses on production readiness:
High Priority -- Required for Launch
| Item | Status | Description |
| Production deployment | In Progress | Apply Terraform modules to AWS environment |
| End-to-end tests | Planned | Playwright browser automation for critical user flows |
| Frontend CI verification | Planned | Ensure npm test and npm run build pass in CI |
| Terraform validation | Planned | Run plan/apply with production AWS credentials |
| Docker Compose smoke test | Planned | Full stack verification (API + workers + DB migrations) |
Medium Priority -- Enhanced MVP
| Item | Status | Description |
| DSR endpoints | Planned | Data export and deletion for GDPR compliance |
| Formal retention policies | Planned | Define data retention by category |
| DPIA assessment | Planned | Data Protection Impact Assessment |
Near-Term Roadmap (Q1-Q2 2026)
| Priority | Item | Business Impact |
| P0 | Production deployment on AWS | Platform goes live for customers |
| P0 | End-to-end testing suite | Ensures critical flows work correctly |
| P1 | BAFA scraper (Germany) | Access to German federal energy/climate programs |
| P1 | KfW scraper (Germany) | Access to Europe's largest promotional bank |
| P1 | Bpifrance scraper (France) | Access to French public investment programs |
| P1 | GDPR DSR implementation | Export/deletion endpoints for compliance |
| P2 | Saved filter presets | Better user experience for repeat searches |
| P2 | Application complexity scoring | Help users prioritize applications |
Medium-Term Roadmap (H2 2026 - H1 2027)
| Priority | Item | Business Impact |
| P1 | Additional national scrapers (RVO, CDTI, IDAE) | Coverage for Netherlands, Spain |
| P1 | Advanced collaborative filtering | Better recommendations from peer behavior patterns |
| P1 | Historical success rate tracking | Data-driven application prioritization |
| P2 | Team collaboration features | Shared watchlists, notes, application co-editing |
| P2 | Advanced analytics dashboard | Deeper insights into funding pipeline and carbon progress |
| P2 | White-label partner portal | Enable partner branding for enterprise distribution |
Long-Term Vision (2027-2030)
| Phase | Timeline | Initiatives |
| Scale | 2027 | 10,000+ paying customers, all 27 EU member states, 500+ active partners |
| Expand | 2028 | International expansion (UK, Switzerland, Norway), mobile PWA, integrated carbon credit marketplace |
| Lead | 2029 | Category leadership in carbon funding, strategic EU institutional partnerships, VCRI global rollout |
| Evolve | 2030+ | Global expansion to emerging markets, platform ecosystem, predictive funding intelligence |
Development Philosophy
How We Build
| Principle | What It Means |
| Test-Driven Development (TDD) | Every feature starts with automated tests that define expected behavior before code is written |
| Carbon-First Design | Every feature decision considers carbon impact; this is not a generic platform with carbon bolted on |
| Accuracy Over Speed | We ship fewer, higher-quality features rather than rushing incomplete functionality |
| Security by Default | Security checks are automated and mandatory; no code ships without passing all gates |
| Infrastructure as Code | Every piece of infrastructure is version-controlled and reproducible |
Release Cadence
- Bug fixes and minor improvements: Continuous deployment (multiple times per week)
- New features: Bi-weekly releases aligned with sprint cycles
- Major platform updates: Quarterly, with advance notice to customers
- Infrastructure changes: Blue/green deployment with zero downtime