InfoLibrarian Knowledge Graph Engine

Spec sheet

Purpose

The engine turns structured and unstructured enterprise metadata into a governed knowledge graph for retrievable, traceable, measurable context.Read 8 Oct 2026 · infolibcorp.com

Intended users

The maker describes the best fit as teams with accumulated document sprawl that need reliable agent context or a data-product foundation, starting with one subject area.Read 8 Oct 2026 · infolibcorp.com

Headquarters

Sarasota, Florida, USARead 28 Sep 2026 · infolibcorp.com

Founded

2005Read 28 Sep 2026 · infolibcorp.com

Commercial availability

The engine is commercial intellectual property available by license, rather than a public self-serve product store.Read 8 Oct 2026 · infolibcorp.com

Corpus hygiene

Ingestion includes hashing, deduplication, near-duplicate detection, version and lineage tracking, and provenance for graph nodes.Read 8 Oct 2026 · infolibcorp.com

Deployment and data control

The stack can run air-gapped or in the cloud; the maker says processing and inference can stay inside the firewall with no third-party model inference.Read 8 Oct 2026 · infolibcorp.com

Document ingestion

It reads presentations, documents, PDFs, spreadsheets, diagrams and audio, preserving table relationships and diagram structure.Read 8 Oct 2026 · infolibcorp.com

Integration options

The platform page names Anthropic, OpenAI, Hugging Face, Ollama, Azure OpenAI, private and on-prem endpoints, and other approved hosts as model options.Read 8 Oct 2026 · infolibcorp.com

Interfaces

The engine provides a command line, programmable API, MCP server and interactive interface for question answering and inspection.Read 8 Oct 2026 · infolibcorp.com

Limitations

The maker says it is not intended as a massive-scale enterprise-wide operational decision platform and describes its ecosystem as small and governance as fit for purpose rather than a formal operating model.Read 8 Oct 2026 · infolibcorp.com

Maker and location

The company describes InfoLibrarian as Brian Brewer’s company and portfolio; its license announcement is datelined Sarasota, Florida.Read 8 Oct 2026 · infolibcorp.com

Measurement

Answers are asynchronously scored for faithfulness, relevancy, contextual precision and contextual recall, with benchmarks per subject area.Read 8 Oct 2026 · infolibcorp.com

Models and infrastructure

It uses a native property graph database and local open-weight models served through Ollama, and is containerized and self-hostable without a managed-service dependency.Read 8 Oct 2026 · infolibcorp.com

Ontology

Users can define and extend node types, relationship types and properties, with extraction constrained to the governed ontology or run schema-free.Read 8 Oct 2026 · infolibcorp.com

Retrieval

Retrieval offers hybrid vector and keyword search, reranking, entity expansion, graph traversal and an explicit deterministic or hybrid mode.Read 8 Oct 2026 · infolibcorp.com

Structured sources

Declarative mappings support CSV, Excel, JSON and XML, alongside enterprise data models, catalogs, process models and organizational structures.Read 8 Oct 2026 · infolibcorp.com

Plans and prices

Every InfoLibrarian Knowledge Graph Engine plan and price →

Where Specifiction places it

• Knowledge Graph Software#23 of 337.5