GraphForge vs. LadybugDB
Which embedded graph engine fits analytical workloads?
LadybugDB is the active continuation of Kuzu, whose repository was archived in October 2025. It is an MIT-licensed, in-process graph database with columnar storage, Cypher, fast bulk loading from Parquet and CSV, and a WebAssembly build. GraphForge is embedded too, so this is the closest architectural comparison in the library. The differences are algorithm breadth, investigation features, and the shape of the results.
On this page
Should I choose LadybugDB or GraphForge?
LadybugDB for WebAssembly or wide language bindings; GraphForge for 94 algorithms and an investigation history.
Choose GraphForge when
- You need more than five algorithms, such as Leiden, Infomap, betweenness, max flow, or Node2Vec, inside the engine.
- Your analysis needs checkpoints, provenance, and recorded competing hypotheses.
- You want every query and verb to return an Apache Arrow table.
Choose LadybugDB when
- You need the graph engine in a browser through WebAssembly, or bindings for Go, Swift, Java, or C.
- You bulk-load large CSV, Parquet, or lakehouse sources such as Iceberg and Delta Lake with COPY FROM.
- You are migrating from Kuzu and want the closest successor.
What can each one do?
19 of 26 capabilities differ. For 12, LadybugDB's docs did not confirm support, so they are marked Not established rather than No.
Each capability has a status and a source. “Not established” means we could not confirm the feature from LadybugDB's primary documentation. It does not mean LadybugDB lacks it.
| Capability | GraphForge | LadybugDB |
|---|---|---|
| Storage | ||
| Property graph model | Shipped Labeled property graph with typed properties. 1 | Yes Property graph data model queried with Cypher. 17 |
| Document model | Not offered No document-store model; properties accept nested, heterogeneous values. 1 | Not established JSON scanning/manipulation is available; no document-store model documented. |
| Relational tables | Not offered Not a relational database; results arrive as Arrow tables. 1 | Partial Can scan attached PostgreSQL, SQLite, DuckDB via extensions; not a SQL engine itself. 21 |
| Key-value access | Not offered No key-value API; data is a property graph. 1 | Not established No key-value API documented. |
| Columnar storage | Shipped Durable projects store graph data as Arrow and Parquet. 4 | Yes Columnar disk-based storage with columnar CSR adjacency indices. 17 |
| Vector search | Shipped Exact cosine retrieval over embeddings you supply; no ANN index. 5 | Extension Official vector extension for similarity search; README calls vector index native. 1721 |
| Full-text search | Shipped Tantivy-backed BM25; hybrid text and vector retrieval through find(). 5 | Extension Official fts extension with BM25 ranking. 21 |
| Geospatial | Not offered No spatial types or indexes are documented. | Not established No spatial extension in official extension list. |
| Execution | ||
| Transactional workloads (OLTP) | Partial Local analytical scope; durable atomic writes with queued and optimistic writer modes. 6 | Partial Serializable ACID transactions, but single READ_WRITE database object per file; optimized for analytics. 1723 |
| Analytical workloads (OLAP) | Shipped Analytical verbs and Cypher over a local columnar project. 7 | Yes Optimized for complex analytical, join-heavy workloads; vectorized, factorized, multi-core execution. 17 |
| Graph traversal | Shipped Cypher MATCH patterns plus BFS, DFS, Dijkstra, A*, and reachability. 8 | Yes Cypher MATCH with recursive relationships and SHORTEST / ALL SHORTEST path syntax. 22 |
| Variable-length paths | Shipped Bounded and unbounded patterns and shortestPath(). 8 | Yes *min..max with WALK/TRAIL/ACYCLIC semantics; default max 30 hops; weighted WSHORTEST. 22 |
| Graph algorithms | Shipped 94 native algorithms behind analysis verbs, not Cypher CALL procedures. 37 | Extension algo extension (PageRank, Louvain, K-Core, WCC, SCC) over projected graphs. 20 |
| Incremental computation | Not offered Algorithms recompute on request; no incremental mode is documented. 7 | Not established No incremental algorithm or view maintenance documented. |
| Deployment | ||
| Embedded / in-process | Shipped In-process library; no server, daemon, or connection string. 9 | Yes 'Embeddable, serverless integration into applications' — runs in-process. 17 |
| Server mode | Planned Optional remote peer extension planned after 0.6.0. 11 | Partial Docs suggest wrapping one READ_WRITE database in an API server (Express.js REST example) for multi-client use. 23 |
| WebAssembly | Not planned A browser WebAssembly engine was declined in ADR 0027. 12 | Yes Ladybug-Wasm (@ladybugdb/wasm-core); default build single-threaded, multithreaded build needs cross-origin isolation. 24 |
| Distributed cluster | Not planned Distributed clusters and HA replication are out of scope. 9 | Not established No clustering or sharding documented; embedded single-process design. |
| Managed cloud | Not planned No hosted compute. The graphforge.sh Hub distributes repositories; execution stays local. 13 | Not established No hosted service found; only cloud-storage extensions (S3, GCS, Azure). |
| Investigation | ||
| Branching | Experimental Research versions, branches, slices, proposals, and forks ship in 0.6.0 release candidates. 14 | Not established No branching feature documented. |
| Snapshots / time travel | Shipped Named checkpoints you can diff and revert, plus transaction-time snapshots. 3 | Not established No as-of / time-travel queries documented. |
| Immutable history | Shipped Knowledge ledger records immutable assertions and evidence lineage. 16 | Not established No immutable version history documented. |
| Provenance | Shipped Provenance and knowledge ledger with evidence lineage. 16 | Not established No provenance tracking documented. |
| Competing hypotheses | Shipped Immutable hypothesis groups with membership and selection events. 3 | Not established |
| Operations | ||
| ACID transactions | Partial Acknowledged-durable atomic publication; docs do not claim general ACID or serializable isolation. 6 | Yes README lists serializable ACID transactions. 17 |
| Replication | Not planned HA replication is out of scope. 9 | Not established No replication documented. |
Showing 26 of 26 capabilities.
How is each one built?
LadybugDB: embedded columnar property graph. GraphForge: an embedded property graph library in Rust, persisted as Parquet.
GraphForge
GraphForge is a Rust library that runs inside your process: no daemon, server, or connection string. Projects persist to a local directory as Arrow and Parquet data plus JSON metadata, and every query and analysis verb returns an Apache Arrow table. Algorithms run in the same native engine as Cypher and are called as API verbs (rank, cluster, paths, analyze, similar), not as Cypher procedures.
LadybugDB
LadybugDB stores node and relationship tables in columnar form on disk with CSR adjacency indexes, and executes Cypher with a vectorized, factorized query processor. Algorithms come from the algo extension and run over connection-scoped projected graphs scanned from disk. One read-write process may open a database at a time.
Both are embedded and columnar. LadybugDB puts query execution first and adds a small algorithm extension. GraphForge puts analysis first: 94 algorithms, BM25 and vector search, and an investigation history share its Arrow and Parquet project.
What the documentation says about LadybugDB
- Kuzu's GitHub repository was archived on 10 October 2025; its README says 'We are archiving the KuzuDB project' and offers v0.11.3 with many bundled extensions.25
- The LadybugDB README states the database was formerly known as Kuzu.17
- v0.12.0 (2025-11-04) was the first release after the rename and described as functionally equivalent to Kuzu v0.11.3.26
- Storage is columnar on disk with columnar sparse row (CSR) adjacency/join indices and a vectorized, factorized query processor.17
- COPY FROM bulk-loads CSV, Parquet, NumPy, DataFrames, JSON and subquery results.27
- Official extensions scan Iceberg, Delta Lake, Unity Catalog, PostgreSQL, SQLite and DuckDB sources.21
- Recursive relationship patterns support WALK, TRAIL and ACYCLIC semantics plus SHORTEST, ALL SHORTEST and weighted WSHORTEST.22
- A GQL query-language extension was added in v0.20.0.28
- Algorithms run on projected graphs that are scoped to a connection and scanned from disk at algorithm time.20
Which one is faster?
Not measured yet. No head-to-head run has been published, so this page shows no performance chart.
Charts appear only after a reproducible run exists: pinned versions, disclosed hardware, generated data, raw results, and scripts.
No comparative results yet
The suites below apply to GraphForge and LadybugDB. Each will be reported separately per benchmark class, never merged into one leaderboard.
Ingestion and writes
| Suite | Workloads | Metrics | Status |
|---|---|---|---|
| INGEST | CSV, JSONL, Parquet, graph edge lists | Records/s, MB/s, total elapsed | Not run |
Query, analytics, and system
| Suite | Workloads | Metrics | Status |
|---|---|---|---|
| LOOKUP | ID lookup, indexed property filter | p50/p95/p99 latency | Not run |
| TRAVERSE | BFS, k-hop, variable-length path | Latency, visited edges/s | Not run |
| JOIN | Multi-relation pattern matching | Latency, rows/s | Not run |
| ALGO | PageRank, WCC, SCC, Louvain, SSSP | Time, memory, correctness | Not run |
| EMBED | Startup, load, reopen, binary size | ms, MiB, disk footprint | Not run |
| DURABLE | Restart, recovery, persistence | Recovery time, data integrity | Not run |
Which graph algorithms ship?
GraphForge: 94 native algorithms. LadybugDB: 8 named in its docs, delivered as extension.
Algorithms are listed by name, as documented. Counts of extension modules or third-party libraries are not presented as native coverage.
GraphForge
94 Rust algorithms run in the engine behind five analysis verbs (rank, cluster, paths, analyze, similar) and return Arrow tables.37
A selection of the catalog:
- PageRank
- ArticleRank
- Betweenness (Brandes)
- Closeness
- Harmonic closeness
- Eigenvector
- HITS
- CELF influence maximization
- Triangle count
- K-core decomposition
- Louvain
- Leiden
- Label propagation
- Infomap
- Walktrap
- Girvan-Newman
- Weakly connected components
- Strongly connected components
- Dijkstra
- A*
- Bellman-Ford
- Yen's k-shortest paths
- Max flow / min cut
- Steiner tree
- Minimum spanning tree
- Topological sort
- Node2Vec
- GraphSAGE
- Fast random projection
- Node similarity (Jaccard)
LadybugDB
Graph algorithms ship in the official algo extension, run via CALL over projected graphs created with PROJECT_GRAPH; shortest paths are native Cypher syntax.2022
Delivered as: extension
- PageRank
- Louvain
- K-Core Decomposition
- Weakly Connected Components
- Strongly Connected Components
- Shortest path (SHORTEST)
- All shortest paths
- Weighted shortest path (WSHORTEST)
Where can each one run?
GraphForge runs in-process only. LadybugDB: embedded yes, server partial, cluster not established.
| Mode | GraphForge | LadybugDB |
|---|---|---|
| Embedded / in-process | Shipped In-process library; no server, daemon, or connection string. 9 | Yes 'Embeddable, serverless integration into applications' — runs in-process. 17 |
| Server mode | Planned Optional remote peer extension planned after 0.6.0. 11 | Partial Docs suggest wrapping one READ_WRITE database in an API server (Express.js REST example) for multi-client use. 23 |
| WebAssembly | Not planned A browser WebAssembly engine was declined in ADR 0027. 12 | Yes Ladybug-Wasm (@ladybugdb/wasm-core); default build single-threaded, multithreaded build needs cross-origin isolation. 24 |
| Distributed cluster | Not planned Distributed clusters and HA replication are out of scope. 9 | Not established No clustering or sharding documented; embedded single-process design. |
| Managed cloud | Not planned No hosted compute. The graphforge.sh Hub distributes repositories; execution stays local. 13 | Not established No hosted service found; only cloud-storage extensions (S3, GCS, Azure). |
| Replication | Not planned HA replication is out of scope. 9 | Not established No replication documented. |
What will my team write?
GraphForge: openCypher from Python, Node.js, Rust. LadybugDB: Cypher, GQL (extension, added v0.20.0).
LadybugDB has the broader binding list, including WebAssembly, and a schema-first Cypher dialect with a GQL extension. Extensions are installed and then loaded per session. GraphForge offers Python, Node.js, and Rust, needs no extension loading, and passes the full openCypher 2024.3 TCK.
- Query languages
- GraphForge openCypher
- LadybugDB Cypher, GQL (extension, added v0.20.0)
- SDKs and bindings
- GraphForge Python, Node.js, Rust
- LadybugDB Python, Node.js, Rust, Go, Swift, Java, C, C++, WebAssembly (JavaScript)
- Written in
- GraphForge Rust
- LadybugDB C++
- Surfaces
- GraphForge Python, Node.js, Rust crates, CLI, VS Code extension, Notebooks, Agent skills
What are the trade-offs?
LadybugDB ships five algorithms and one writer process. GraphForge has no WASM build and fewer bindings.
A fair assessment cuts both ways.
GraphForge
- No WebAssembly or browser build; it was declined in ADR 0027.
- No native CSV or Parquet COPY FROM loader in stable 0.5.2; Parquet import sessions are experimental in 0.6.0.
- Fewer language bindings (Python, Node.js, Rust).
LadybugDB
- Only one READ_WRITE Database object may open a database; other processes can only open it READ_ONLY.23
- Extensions must be installed and then loaded per session; they are scoped to the session that loads them.21
- Projected graphs are connection-bound, a documented known limitation with the Python async connection pool ('Cannot find graph').20
- The default WebAssembly build has no multi-threading, which the docs note may hurt performance.24
- Multi-labelled Neo4j nodes cannot be migrated because Ladybug does not support them yet.19
How would I migrate?
Load LadybugDB data through Arrow, pandas, or Polars. Standard openCypher carries over; procedure calls become verb calls.
Into GraphForge
- GraphForge 0.5.x loads data through Arrow, pandas, and Polars bulk calls or Cypher scripts; it has no native CSV loader.15
- Algorithms are API verbs (rank, cluster, paths, analyze, similar), not Cypher CALL procedures, so procedure calls become verb calls.7
- GraphForge passes the full openCypher 2024.3 TCK, so standard openCypher queries are the portable part of a migration.10
Moving from LadybugDB
- Existing Kuzu users: v0.12.0 is described as functionally equivalent to Kuzu v0.11.3 apart from the name change.26
- The neo4j extension's NEO4J_MIGRATE pulls data from Neo4j, requiring APOC file export to be enabled.19
- Cypher is the primary query language, easing porting from other Cypher databases, though label semantics differ (no multi-label nodes).19
What else do people ask about LadybugDB?
What happened to Kuzu? Kuzu was archived in October 2025; LadybugDB continues it; GraphForge is an embedded alternative with more algorithms.
What happened to Kuzu?
Kuzu's GitHub repository was archived on 10 October 2025. LadybugDB continues the codebase: its first release, v0.12.0, was described as functionally equivalent to Kuzu v0.11.3.
Is GraphForge a Kuzu alternative?
Yes, for embedded graph analysis. Both run in-process with Cypher. GraphForge has a much larger algorithm catalog and investigation features; LadybugDB has WebAssembly and more language bindings.
Which algorithms does LadybugDB ship?
The algo extension provides PageRank, Louvain, k-core decomposition, and weakly and strongly connected components. Shortest, all-shortest, and weighted shortest paths are native Cypher syntax.
How was this researched?
From 28 primary sources, accessed 2026-10-19, each cited where it is used.
Capabilities were checked against primary documentation, repositories, and release notes, accessed 2026-10-19. Each cell records two things: the capability’s state and the evidence behind it. Every GraphForge cell carries a release status. Shipped means stable 0.5.2; experimental means 0.6.0 release candidates only. GraphForge is pre-1.0 and makes no backward-compatibility guarantee yet.
- Yes documented natively
- Partial limited or caveated
- Extension official extension or module
- Paid edition paid or enterprise edition only
- Removed removed, with the version cited
- Not established not confirmed by our research; not a “no”
- Shipped in stable 0.5.2
- Partial shipped with documented limits
- Experimental in 0.6.0 release candidates only
- Planned on the roadmap
- Not offered outside the product's model
- Not planned explicitly out of scope
Benchmark classes
- Equivalent workloads. The same logical operation with the same correctness requirement on every engine.
- Native optimized workloads. Each engine's best supported implementation, including projections, indexes, extensions, and native algorithms.
- End-to-end application workloads. Time from raw data to a usable result: import, indexing, projection, computation, and output retrieval.
Sources · accessed 2026-10-19
- GraphForge repository README — GitHub
- GraphForge releases — GitHub
- GraphForge 0.5.0 changelog — GitHub
- Storage architecture — GraphForge Book
- Analytics integration: search and retrieval — GraphForge Guide
- ADR 0018: acknowledged durability and isolation — GitHub
- Algorithm verbs — GraphForge Book
- Cypher guide — GraphForge Guide
- openCypher compatibility — GraphForge Reference
- openCypher TCK compliance — GraphForge Reference
- Remote authority peer extension (#215) — GitHub
- ADR 0027: native runtime boundary — GitHub
- Release scope and roadmap — GraphForge Docs
- Keep research history — GraphForge Guide
- Graph construction — GraphForge Guide
- Knowledge public API v1 — GraphForge Book
- LadybugDB/ladybug README — GitHub
- Releases — LadybugDB/ladybug on GitHub
- Neo4j extension — Ladybug docs
- Graph algorithms (algo extension) — Ladybug docs
- Extensions — Ladybug docs
- MATCH — Ladybug Cypher manual
- Connections and concurrency — Ladybug docs
- Ladybug-Wasm — Ladybug docs
- kuzudb/kuzu (archived) README — GitHub
- Release v0.12.0 — LadybugDB/ladybug on GitHub
- Import data — Ladybug docs
- Release v0.20.0 — LadybugDB/ladybug on GitHub
Try GraphForge on your own graph.
Install the package, open a project, and run the same notebook we measure with.