Skip to content

Roadmap

This roadmap outlines the planned development of Grafeo. Priorities may shift based on community feedback and real-world usage.

0.1.x - Foundation

Building a fully functional graph database

Core Database

  • Labeled Property Graph (LPG) storage model
  • MVCC transactions with snapshot isolation
  • Multiple index types (hash, B-tree, trie, adjacency)
  • Write-ahead logging (WAL) for durability
  • In-memory and persistent storage modes

Query Languages

  • GQL (ISO standard) - full support
  • Cypher - experimental
  • Gremlin - experimental
  • GraphQL - experimental
  • SPARQL - experimental

Data Models

  • LPG - full support
  • RDF - experimental

Bindings

  • Python - full support via PyO3
  • NetworkX integration - experimental
  • solvOR graph algorithms - experimental

0.2.x - Performance

Competitive with the fastest graph databases

Performance Improvements

  • Factorized query processing for multi-hop traversals (avoid Cartesian products)
  • Worst-case optimal joins via Leapfrog TrieJoin for cyclic patterns (O(N^1.5) triangles)
  • Lock-free concurrent reads using DashMap-backed hash indexes (4-6x improvement)
  • Direct lookup APIs bypassing query planning for O(1) point reads (10-20x faster)
  • Query plan caching with LRU cache for repeated queries (5-10x speedup)
  • NUMA-aware scheduling with same-node work-stealing preference

New Features

  • Ring Index for RDF (ring-index feature) - 3x space reduction using wavelet trees
  • Block-STM parallel execution (block-stm feature) - optimistic parallel transactions
  • Tiered hot/cold storage (tiered-storage feature) - compressed epoch archival
  • Succinct data structures (succinct-indexes feature) - rank/select bitvectors, Elias-Fano

Expanded Support

  • RDF - full support with Ring Index and SPARQL optimization
  • All 5 query languages promoted to full support
  • NetworkX and solvOR integrations promoted to full support

0.3.x - AI Compatibility

Ready for modern AI/ML workloads

Vector Features

  • Vector Type - First-class Vector type with f32 storage (4x compression vs f64)
  • Distance Functions - Cosine, Euclidean, Dot Product, Manhattan metrics
  • HNSW Index - O(log n) approximate nearest neighbor search with batch insert/search
  • Hybrid Queries - Combine graph traversal with vector similarity in GQL/Cypher/SPARQL
  • Serializable Isolation - SSI for write skew prevention and strong consistency

Vector Quantization

  • Scalar Quantization - f32 → u8, 4x compression with ~97% recall
  • Binary Quantization - f32 → 1 bit, 32x compression with SIMD-accelerated hamming distance
  • Product Quantization - Codebook-based 8-32x compression with asymmetric distance computation
  • QuantizedHnswIndex - Two-phase search with rescoring support
  • SIMD Acceleration - AVX2+FMA, SSE, and NEON support for 4-8x faster distance computations
  • Memory-Mapped Vector Storage - Disk-backed storage with LRU cache for large datasets
  • VectorScan Operators - HNSW and brute-force search in query execution plans
  • VectorJoin Operator - Hybrid graph pattern + vector similarity search
  • Vector Zone Maps - Centroid and bounding box pruning for block skipping
  • Vector Cost Estimation - HNSW O(ef * log N) and brute-force O(N) cost models
  • Python Quantization API - Full quantization support from Python

Execution & Quality

  • Selective Property Loading - Projection pushdown for O(N*K) vs O(N*C) reads
  • Parallel Node Scan - Morsel-driven parallel execution for 3-8x speedup on large scans
  • Query Performance Metrics - Execution timing and row counts on query results
  • Error Message Suggestions - Fuzzy "Did you mean X?" hints for undefined variables and labels
  • Adaptive WAL Flusher - Self-tuning background flush with consistent cadence

Syntax Support

-- Vector literals and similarity functions
MATCH (m:Movie)
WHERE cosine_similarity(m.embedding, $query) > 0.8
RETURN m.title

-- Create vector index
CREATE VECTOR INDEX movie_embeddings ON :Movie(embedding)
  WITH (dimensions: 384, metric: 'cosine')

0.4.x - Developer Accessibility

Reach more developers

New Bindings (Experimental)

  • Node.js / TypeScript (@grafeo-db/js) - native bindings with full type definitions
  • WebAssembly (@grafeo-db/wasm) - raw WASM binary for edge runtimes
  • Go - CGO bindings for cloud-native applications

Ecosystem Integration

  • grafeo-server - standalone HTTP server with web UI, Docker image
  • grafeo-web - Grafeo in the browser via IndexedDB, Web Workers, React/Vue/Svelte

Query Languages

  • SQL/PGQ (SQL:2023) - GRAPH_TABLE function for SQL-native graph queries

Quality

  • Continued bug fixes
  • Stricter linting rules
  • Performance tuning based on real-world usage

0.5.x - Beta

Preparing for production readiness

Focus Areas

  • Performance optimizations across all components
  • Lots of bug hunting and fixing
  • Documentation improvements, user guides and examples
  • API stabilization

Goal

  • Ready for production evaluation
  • Clear path to 1.0

Future Considerations

These features are under consideration for future releases:

  • Additional language bindings (Java/Kotlin, Swift)
  • Distributed/clustered deployment
  • Cloud-native integrations

Contributing

Interested in contributing to a specific feature? Check our GitHub Issues or join the discussion.

Last updated: February 2026