What is this?
Semantica is an open-source, graph-native infrastructure layer that ingests enterprise data and builds a structured Context Graph and knowledge graph with deterministic reasoning, ontology governance, and full decision provenance, so AI systems can be explained and audited instead of treated as black boxes.
Semantica is an open-source, graph-native infrastructure layer that ingests enterprise data and builds a structured Context Graph and knowledge graph with deterministic reasoning, ontology governance, and full decision provenance, so AI systems can be explained and audited instead of treated as black boxes.Is this a framework, or just another vector database?It is a modular Python framework built as a full ingest-to-export pipeline: ingest, extract, kg, reasoning, ontology, provenance and export are independently importable modules, plus a CLI, REST API, and MCP server on top, rather than a single library with one feature.
It is a modular Python framework built as a full ingest-to-export pipeline: ingest, extract, kg, reasoning, ontology, provenance and export are independently importable modules, plus a CLI, REST API, and MCP server on top, rather than a single library with one feature.How do I get it running?You install the semantica Python package and run the built-in Quick Start check, which verifies the install in about 5 seconds; from there you wire up pipelines with the declarative semantica.pipeline DSL covering ingest, extract, build, and export.
You install the semantica Python package and run the built-in Quick Start check, which verifies the install in about 5 seconds; from there you wire up pipelines with the declarative semantica.pipeline DSL covering ingest, extract, build, and export.Do I need an LLM or API key to use it?No, graph construction, reasoning, and provenance are fully deterministic and run without any LLM; an LLM is optional, and when used, any major provider (OpenAI, Anthropic, Gemini, etc.) can plug in through semantica.llms.
No, graph construction, reasoning, and provenance are fully deterministic and run without any LLM; an LLM is optional, and when used, any major provider (OpenAI, Anthropic, Gemini, etc.) can plug in through semantica.llms.Can it explain what's happening inside the LLM itself?No, Semantica only explains what's outside the model: the context fed in, the decision produced, its provenance, and the policies applied; the LLM's own internal reasoning stays opaque, just like with any other external system.
No, Semantica only explains what's outside the model: the context fed in, the decision produced, its provenance, and the policies applied; the LLM's own internal reasoning stays opaque, just like with any other external system.Who is this actually built for?It targets AI/ML platform teams shipping decision-making agents, enterprise data teams on Databricks, Snowflake or SAP, compliance and audit teams, and regulated industries like finance, healthcare, legal, and government that can't ship a black box or hand data to a third-party SaaS.
It targets AI/ML platform teams shipping decision-making agents, enterprise data teams on Databricks, Snowflake or SAP, compliance and audit teams, and regulated industries like finance, healthcare, legal, and government that can't ship a black box or hand data to a third-party SaaS.Does my enterprise data leave my own systems?No, it's self-hosted and open source with zero vendor lock-in; data already in your Databricks, Snowflake or SAP lakehouse/warehouse becomes graph nodes in place, without an export/import hop to a third-party SaaS.
No, it's self-hosted and open source with zero vendor lock-in; data already in your Databricks, Snowflake or SAP lakehouse/warehouse becomes graph nodes in place, without an export/import hop to a third-party SaaS.What makes a 'decision' a first-class object instead of just a log line?Calling recorddecision() creates a permanent, structured graph node with a full lifecycle, exportable as W3C PROV-O, so it's queryable by precedent and causally linked, which is the format most compliance frameworks accept for regulator submission.
Calling recorddecision() creates a permanent, structured graph node with a full lifecycle, exportable as W3C PROV-O, so it's queryable by precedent and causally linked, which is the format most compliance frameworks accept for regulator submission.Which graph databases and vector stores can I plug in?Semantica is polyglot: RDF triple stores (Oxigraph, Blazegraph, Jena, RDF4J) and Labeled Property Graphs (Neo4j, FalkorDB, AGE, Neptune), plus vector stores like FAISS, Qdrant, Weaviate, Milvus, Pinecone, and PgVector, all swappable without touching your code.
Semantica is polyglot: RDF triple stores (Oxigraph, Blazegraph, Jena, RDF4J) and Labeled Property Graphs (Neo4j, FalkorDB, AGE, Neptune), plus vector stores like FAISS, Qdrant, Weaviate, Milvus, Pinecone, and PgVector, all swappable without touching your code.Does it work with agent frameworks like LangChain?Yes, it has drop-in integrations for Agno, CrewAI, and LangChain, plus a full MCP server, CLI, REST API, and plugins across major editors.
Yes, it has drop-in integrations for Agno, CrewAI, and LangChain, plus a full MCP server, CLI, REST API, and plugins across major editors.What happens when two data sources disagree on a fact?Conflicting facts are flagged and resolved during the conflict-detection and deduplication stage rather than silently overwritten, so duplicates get merged and contradictions surface instead of corrupting the knowledge graph.
Conflicting facts are flagged and resolved during the conflict-detection and deduplication stage rather than silently overwritten, so duplicates get merged and contradictions surface instead of corrupting the knowledge graph.Which topics is it in?ontology, explainable-ai, ai-infrastructure, ai-governance, neo4j, graph-database, reasoning, retrieval-augmented-generation, data-engineering, knowledge-graph, context-engineering, vector-database
ontology, explainable-ai, ai-infrastructure, ai-governance, neo4j, graph-database, reasoning, retrieval-augmented-generation, data-engineering, knowledge-graph, context-engineering, vector-database