Tentris

Tentris is a self-hosted RDF database for storing and querying knowledge graphs with SPARQL 1.1. It is most suitable for developers, students and researchers building non-commercial semantic projects, while commercial teams need the paid licensing terms. Its compact index and streaming query engine are compelling, but the free edition is tightly limited for production use.

Overview

Tentris stores RDF graphs in a single compressed Hypertrie index intended to reduce memory consumption. That design is relevant when you operate the database in your own infrastructure and want the index to use resources efficiently, although the Community plan keeps storage in memory and therefore does not address persistence requirements.

The engine combines adaptive query planning with worst-case optimal joins. It streams results and avoids materializing intermediate results, an approach aimed at keeping complex graph queries from creating unnecessarily large temporary datasets. RDF 1.1 and SPARQL 1.1 compliance means an existing RDF data model can be retained when moving to Tentris. Federated queries are supported, but reasoning is not.

Tentris is also positioned as a knowledge layer for GraphRAG systems, AI agents and other semantic applications. Its stated goal is deterministic retrieval across connected knowledge rather than an opaque retrieval layer. The company ties published benchmark figures to reproducible datasets, queries, hardware and methods, giving technical buyers a way to compare those results with their own workloads. The product grew from more than a decade of research at Paderborn University, and Tentris says it is built in Germany.

Key features

Storage and query execution

The compressed single-index model, streaming execution and avoidance of intermediate-result materialization are the product’s clearest technical differentiators. They should appeal to teams concerned about memory behavior in graph workloads, but they do not remove the need to size infrastructure for the actual dataset and query mix.

Standards and formats

Tentris supports SPARQL 1.1 and RDF 1.1, plus N-Triples, Turtle, N-Quads, TriG and RDF/XML. Support for common serializations lowers migration friction. Federated queries extend access across graph sources; the absence of reasoning makes it less suitable when inferred facts are a core requirement.

Deployment and data control

Linux, macOS and self-hosted deployment are supported, with Docker and Python installation options. Builds are listed for x86-64 and ARM AARCH64. The downloadable software runs inside the licensee’s infrastructure, and Tentris says it does not host, receive or otherwise access the content. A planned cloud service for Azure, AWS and Google Cloud is marked coming soon, so it is not a current deployment choice.

Pricing

Community is the practical entry point for developers, students and researchers evaluating or building private, academic, educational, scientific or research projects. Its single-core, in-memory configuration is a meaningful cap, and its EULA excludes production and revenue-generating use. Tentris provides community support with the plan, while the EULA says it provides no support or maintenance for the non-commercial software, so organizations should not treat the free edition as a supported production service.

Commercial is the relevant route for production workloads, persistence and formal operational commitments. The 12-core starting point and annual, per-core licensing can make it disproportionate for a small deployment, but it adds persistent storage, ACID transactions and an optional SLA. The 60-day trial license is longer than a short evaluation window. Cloud is for teams that prefer managed consumption, backups and scaling, but it remains an announced future offering rather than an immediately usable plan.

Platforms

Tentris runs on Linux and macOS and is designed for self-hosting. Docker and Python installation paths are listed, alongside x86-64 and ARM AARCH64 support. Web access is listed as a platform, while the planned cloud service has not yet launched.

Who it's for

Choose Tentris Community for non-commercial graph experiments, coursework, research and semantic prototypes where in-memory operation and one CPU core are sufficient. It is also a sensible candidate for GraphRAG or agent experiments that need deterministic SPARQL retrieval without sending data to a hosted provider.

Choose Commercial when the database must serve production or revenue-generating applications, retain data persistently, use ACID transactions or obtain an optional SLA. Look elsewhere if rule-based reasoning is essential, if a small commercial installation cannot justify a 12-core starting point, or if you need a cloud service today.

Pros and cons

• Pros: compressed indexing targets lower memory use; adaptive, worst-case optimal execution streams results without intermediate materialization; RDF 1.1 and SPARQL 1.1 support eases migration; federated queries and broad RDF format support help mixed graph environments; self-hosting keeps content within the licensee’s infrastructure.
• Cons: Community is single-core and in-memory; its license excludes production and revenue-generating use; the EULA provides no support or maintenance for the non-commercial software; reasoning is not supported; commercial licensing starts at 12 CPU cores and is custom-priced; cloud deployment is not yet available.

Alternatives

Eclipse RDF4J is a free open-source Java framework if you want a framework rather than Tentris’s packaged database approach; its latest release requires Java 25 or newer.

Fluree DB is a free, self-hostable option with native JSON-LD and SPARQL, including an open-source edition.

GraphDB is worth considering for a broader platform with a free tier, though GraphDB Free allows one parallel query and five repositories.

QLever is a free Apache 2.0-licensed alternative for teams that prefer open-source software.

RDFox fits buyers seeking a commercial product with a free evaluation license and a 30-day trial.

Stardog is another freemium choice, with a renewable one-year free license, up to one million stored edges and three databases.

Apache Jena Fuseki is a free Apache-licensed self-hosted server.

Parliament is a free BSD 3-Clause open-source RDF store for Linux, macOS and Windows.

Browse the RDF Databases directory for additional choices.

Verdict

Tentris is a strong fit for non-commercial users who value self-hosting, standards compatibility and memory-conscious SPARQL execution, especially in research and semantic-AI prototypes. Its main reason to choose is the combination of compressed indexing and streaming, adaptive query processing. Its main reason to look elsewhere is the licensing boundary: production users face custom annual pricing from 12 cores, while the free plan cannot be used for revenue-generating work and does not provide persistent storage.