Roadmap

This roadmap outlines the current state of ContextRouter and planned enhancements. We use NotImplementedError stubs throughout the codebase to mark areas ready for contribution.

Current Status (v0.10.x)

Stable Features ✅

Feature	Status	Notes
LangGraph orchestration	✅ Stable	Cortex with customizable graphs
Multi-provider LLMs	✅ Stable	Vertex, OpenAI, Anthropic, Groq, Local
PostgreSQL + pgvector	✅ Stable	Hybrid search with FTS
Staged ingestion pipeline	✅ Stable	Preprocess → Structure → Index → Deploy
CLI tools	✅ Stable	Full pipeline management
Bisquit provenance	✅ Stable	Data tracking throughout

Beta Features 🔶

Feature	Status	Notes
Knowledge Graph (Cognee)	🔶 Beta	Entity extraction, graph building
Vertex AI Ranking	🔶 Beta	Neural reranking
AG-UI protocol	🔶 Beta	Streaming to frontends
OpenRouter support	🔶 Beta	Hundreds of models

Alpha Features 🔷

Feature	Status	Notes
HuggingFace local	🔷 Alpha	In-process inference
Plugin auto-discovery	🔷 Alpha	Automatic loading
A2A protocol	🔷 Alpha	Multi-agent communication

Near-Term Priorities

Storage & Retrieval

• Postgres hybrid search improvements — Better ltree taxonomy filtering
• Local cross-encoder reranking — FlashRank or sentence-transformers
• Query expansion — Automatic query reformulation
• Incremental indexing — Update without full re-ingestion

Model Support

• Streaming improvements — Better handling of partial tokens
• Multimodal embeddings — Image + text embeddings
• Cost tracking — Per-request cost estimation

New Connectors

• Notion — Pull from Notion workspaces
• Slack — Search Slack history
• GitHub — Index repository content
• Confluence — Enterprise wiki integration

Mid-Term Goals

Developer Experience

• Playground UI — Web-based testing interface
• VS Code extension — Inline documentation and snippets
• Better error messages — Actionable suggestions
• Type stubs — Full type coverage for IDE support

RAG Improvements

• Multi-hop reasoning — Follow-up queries for complex questions
• Confidence scoring — Indicate certainty of answers
• Source highlighting — Show exactly which text was used
• Answer comparison — Multiple candidate responses

Ingestion Enhancements

• Parallel processing — Multi-threaded chunk processing
• Duplicate detection — Cross-document deduplication
• Change detection — Only re-process modified sections
• Format expansion — DOCX, PPTX, HTML support

Long-Term Vision

Enterprise Features

Feature	Description
Multi-tenancy	Isolated knowledge bases per tenant
RBAC	Role-based access control
Audit logging	Full operation history
SSO integration	SAML, OIDC support
Data residency	Region-specific storage

Advanced Capabilities

Feature	Description
Multi-agent orchestration	Coordinated specialist agents
Autonomous planning	Break down complex tasks
Tool use	Function calling, API integrations
Memory systems	Long-term conversation memory
Self-improvement	Learn from feedback

Contributing

We welcome contributions! Areas marked with NotImplementedError are excellent starting points.

High-Impact Areas

modules/connectors/      → New data sources
modules/providers/       → New storage backends
modules/transformers/    → New data enrichment
cortex/nodes/           → New graph capabilities

Getting Started

1. Fork the repository
2. Find a NotImplementedError stub
3. Implement the feature
4. Add tests
5. Submit a PR

See CONTRIBUTING.md for detailed guidelines.

Release Cycle

Type	Frequency	Examples
Patch (0.10.x)	Weekly	Bug fixes, docs
Minor (0.x.0)	Monthly	New features
Major (x.0.0)	As needed	Breaking changes

Feedback

• Feature requests — GitHub Issues
• Discussions — GitHub Discussions
• Security issues — security@contextrouter.dev