Fluree DB

Fluree DB is a graph database built around data that needs to be trusted and audited over time: every fact is temporal and verifiable, and the database supports git-like branching and merging of data itself, not just schema migrations. It’s standards-compliant across RDF 1.1/1.2, SPARQL, JSON-LD, and openCypher, so teams aren’t locked into a Fluree-specific query language.

Benchmarks published by the team claim over 2 million facts/second bulk import and completion of all 850 WGPB graph-pattern queries against the full 21.5-billion-triple Wikidata dump with a 43ms geometric mean — reported as 10.4x faster than the next database on that same benchmark. Vector, text, and geo search are integrated directly rather than requiring separate systems, along with fine-grained access control with no external dependencies.

Fluree DB is licensed under the Business Source License 1.1 (BUSL), not a standard OSI-approved open-source license: source is available and self-hosting/internal use is fully permitted, but you may not offer it as a hosted “Database Service” to third parties. Each version converts to Apache-2.0 four years after its release. A related feature, Fluree Memory, gives AI coding assistants like Claude Code and Cursor persistent, git-shareable project memory backed by a Fluree ledger.

What You Get

• A temporal, verifiable graph database supporting RDF, SPARQL, JSON-LD, and openCypher queries
• Git-like branching and merging of data itself, not just schema changes
• Integrated vector, text, and geo search without requiring separate systems
• Fluree Memory, a persistent project-memory feature for AI coding assistants backed by a Fluree ledger

Common Use Cases

• Running a graph database where data provenance and historical verifiability matter, not just current state
• Branching and merging large datasets the way you’d branch and merge code, for exploratory or multi-team data work
• Combining vector search, full-text search, and geo queries against graph data in one database instead of three systems
• Giving AI coding assistants persistent, git-shareable project memory via the Fluree Memory feature

Under The Hood

Architecture Fluree DB’s temporal model treats history as a first-class feature rather than an add-on audit log, and its git-like branching/merging extends version-control semantics to the data layer itself — letting teams branch a dataset, make changes, and merge them back the way they would with code. Standards compliance (RDF, SPARQL, JSON-LD, openCypher) is built in rather than requiring translation layers, and vector/text/geo search are integrated directly into the core query engine instead of bolted on via external services.

Tech Stack Rust for the core engine, with deployment options spanning a free serverless cloud tier (usage-limited), Docker for self-hosted deployment, and a local CLI. Benchmarks are run and published against large-scale public datasets (the full Wikidata dump) rather than only synthetic test data.

Code Quality Very active, consistently maintained commit history and a strong watcher-to-star engagement ratio (per health metrics) suggest a serious, technically engaged user base rather than passive stargazing; published, reproducible benchmarks against a well-known public dataset add credibility beyond marketing claims.

What Makes It Unique Most graph databases pick one of standards-compliance, verifiable history, or integrated multi-modal search; Fluree DB combines all three — RDF/SPARQL/openCypher standards support, git-like temporal branching, and integrated vector/text/geo search — in a single engine, backed by benchmarks against one of the largest public graph datasets available.

Self-Hosting

Licensing Model Business Source License 1.1 (BUSL) — not a standard OSI-approved open-source license. Source code is publicly available and self-hosting/internal use is fully permitted, including as a component of your own application; the specific restriction is that you may not offer the Licensed Work as a hosted “Database Service” to third parties.

Self-Hosting Restrictions Self-hosting for internal use is explicitly permitted with no functional restrictions. The only prohibited use case is reselling Fluree DB itself as a managed database service to third parties.

Cloud vs Self-Hosted A free, usage-limited serverless cloud tier is available at flur.ee/solo, alongside self-hosted Docker deployment.

License Key Required No, for self-hosted/internal use. Each version’s Change Date converts it to Apache-2.0 four years after its first public release.