Semantica is an open-source Python framework for building context graphs and decision intelligence layers on top of AI agents and LLM applications. It provides structured context, causal reasoning, and full decision provenance, making AI explainable and traceable.

Yes. The core library is 100% free and open source under the MIT license. No feature gating, no usage limits, no hidden tiers.

How is Semantica different from LangChain or LlamaIndex?

LangChain and LlamaIndex focus on orchestration and retrieval. Semantica adds a structured knowledge layer — context graphs, causal decision tracking, and W3C PROV-O provenance — that integrates on top of any framework.

Which AI frameworks does Semantica support?

Semantica integrates with LangGraph, CrewAI, LlamaIndex, Claude Code, AutoGen, and any Python-based agent framework, along with 100+ LLM models via LiteLLM.

What is a context graph?

A context graph is a structured, queryable knowledge graph that stores entities, relationships, and decisions made by your AI — with temporal validity and full provenance. Unlike a vector store, it understands causal relationships.

Does Semantica require a vector database?

No. Semantica works standalone. When you want hybrid graph + semantic retrieval it integrates natively with FAISS, Pinecone, Weaviate, Qdrant, Milvus, and pgvector.

What is decision provenance?

Decision provenance means every action your AI takes is stored as a first-class object with its inputs, reasoning chain, and outcome — compliant with W3C PROV-O. You can always audit why the AI did what it did.

How do I get started?

Run pip install semantica, then follow the quickstart in the docs at docs.getsemantica.ai. The core graph is up and running in under 10 lines of Python.

Integrations

Plugs into the stack you already run

Warehouses, graph stores, vector databases, LLM providers, agent frameworks, and AI editors. Every integration below ships with Semantica.

/

DatabricksSnowflakeSAPSalesforceGoogle BigQueryAmazon RedshiftMicrosoft Power BIApache AirflowNeo4jFalkorDBAWS NeptuneOxigraphQdrantPineconeWeaviateMilvusDatabricksSnowflakeSAPSalesforceGoogle BigQueryAmazon RedshiftMicrosoft Power BIApache AirflowNeo4jFalkorDBAWS NeptuneOxigraphQdrantPineconeWeaviateMilvus

Claude CodeCursorCodex CLIWindsurfClineVS CodeAnthropicGoogle GeminiOllamaDeepSeekCrewAILangChainGoogle ADKAgnoOpenAIClaude CodeCursorCodex CLIWindsurfClineVS CodeAnthropicGoogle GeminiOllamaDeepSeekCrewAILangChainGoogle ADKAgnoOpenAI

Four ways in

Connect it your way

Plugin bundles

Skills, agents, and hooks for Claude Code, Cursor, Codex CLI, and pi.

MCP server

Any MCP-compatible client connects in about 30 seconds.

python -m semantica.mcp_server

REST API

A full REST backend, ready for any HTTP client or agent.

Python SDK

Import only the modules you need, from a single package.

pip install semantica

Recently added

New integrations, release by release

Full changelog

v0.7.0

Sep 22, 2026

v0.6.8

Sep 5, 2026

v0.6.7

Aug 28, 2026

v0.6.6

Aug 20, 2026

v0.6.5

Aug 11, 2026

v0.6.0

Jul 21, 2026

82 of 82

Enterprise Data Platforms

Pull tables and metadata straight from your lakehouse, warehouse, ERP, CRM, or BI layer, with lineage, instead of exporting to CSV first.

Databricks

New v0.6.0

Unity Catalog and Delta Lake, PAT or OAuth M2M, catalog and lineage introspection

Delta Lake

Snowflake

Warehouse, database and schema access with password, key-pair, or OAuth auth

Warehouse

SAP

New v0.6.7

OAuth2 or Basic auth, Business Partners and Sales Orders

OData v2/v4

Salesforce

New v0.6.8

CRM ingestion with schema discovery, custom objects, and sandbox support

REST · SOQL

Google BigQuery

New v0.7.0

Table and query ingestion with ADC or service-account key auth

Warehouse

Amazon Redshift

New v0.7.0

Table and query ingestion with password or IAM authentication

Warehouse

Microsoft Power BI

New v0.7.0

Workspace, dataset, report, and dataflow metadata via Azure AD OAuth2

REST

Apache Airflow

New v0.7.0

DAG and task-run metadata ingestion via the Airflow REST API

REST

AI Coding Assistants

Native plugin bundles, an MCP server, or a REST API. Semantica meets your editor where it already lives.

Claude Code

Skills, agents, and hooks

Plugin bundle

Cursor

Skills and agents

Plugin bundle

Codex CLI

Skills and agents

Plugin bundle

pi

Skills and plugin

Plugin bundle

Windsurf

Auto tool connection over MCP

MCP + plugin

Cline

Auto tool connection over MCP

MCP + plugin

Continue

Auto tool connection over MCP

MCP + plugin

VS Code

Auto tool connection over MCP

MCP + plugin

OpenClaw

MCP server and dedicated plugin

MCP + plugin

Claude Desktop

Standalone MCP server for desktop

MCP server

GitHub Copilot

Full REST backend

REST API

Roo Code

Full REST backend

REST API

Goose

Full REST backend

REST API

Kilo Code

Full REST backend

REST API

Aider

Full REST backend

REST API

Amazon Q

Full REST backend

REST API

Zed

Full REST backend

REST API

Agentic Frameworks

First-class toolkits for the frameworks teams ship agents on. Everything else connects through the REST API and MCP today.

Agno

Decision intelligence, context graphs, and GraphRAG toolkits

Native

CrewAI

New v0.6.6

First-class integration via semantica[crewai]

Native

LangChain

New v0.6.7

First-class integration via semantica[langchain]

Native

Google ADK

New v0.6.8

Agent Development Kit integration

Native

LangGraph

Connect stateful graph agents over REST API and MCP

REST + MCP

LlamaIndex

Connect over REST API and MCP

REST + MCP

AutoGen

Connect over REST API and MCP

REST + MCP

OpenAI Agents SDK

Connect over REST API and MCP

REST + MCP

LLM Providers

Bring any model. First-class wrappers for the major providers, and everything else routed through LiteLLM.

Anthropic

New v0.6.8

First-class provider wrapper

Claude

Google Gemini

New v0.6.8

First-class provider wrapper

Gemini

Ollama

New v0.6.8

Run open models locally, first-class wrapper

Local

DeepSeek

New v0.6.8

First-class provider wrapper

Hosted

Novita

New v0.6.8

First-class provider wrapper

Hosted

OpenAI

GPT-4o, o1, and o3 through the native wrapper

GPT · o-series

Groq

First-class provider wrapper

Hosted

Hugging Face

Models from the Hugging Face Hub

Hub

Mistral

Routed through LiteLLM

via LiteLLM

Cohere

Routed through LiteLLM

via LiteLLM

AWS Bedrock

Routed through LiteLLM

via LiteLLM

LiteLLM

Azure OpenAI, Llama, Perplexity, Together AI, Fireworks AI, Replicate, and more

Gateway

Graph Databases

Labeled property graphs behind one unified interface. Swap backends with configuration, not code.

Neo4j

Industry-standard graph database with full Cypher support

Cypher

FalkorDB

Native support, high-performance graphs

Cypher

Apache AGE

PostgreSQL extension with openCypher queries

openCypher

AWS Neptune

IAM authentication and full Gremlin support

Gremlin

RDF Triple Stores

SPARQL over HTTP or embedded, with bulk loading and CONSTRUCT query templates across every backend.

Oxigraph

New v0.6.5

In-memory or on-disk, no server required

Embedded

Altair Anzo

New v0.6.6

Enterprise knowledge graph platform backend

SPARQL

Blazegraph

High-performance RDF triple store

SPARQL

Apache Jena

Full-featured RDF framework

SPARQL

Eclipse RDF4J

Java framework for RDF and SPARQL

SPARQL

Vector Stores

Semantic and hybrid search on your preferred backend. Every store supports enumeration with scan_vectors() as of v0.6.8, so you can migrate between them.

FAISS

Meta's dense similarity search

Local

Qdrant

High-performance vector database

Self-host

Weaviate

AI-native search and vector store

Self-host

Milvus

Scalable similarity search

Distributed

Pinecone

Managed vector database

Managed

PgVector

PostgreSQL extension for embeddings

Postgres

SQLite-vec

New v0.6.0

Lightweight embedded vector storage

Embedded

Data Sources & Connectors

One ingestion interface for files, warehouses, databases, streams, and repositories.

DuckDB

Query local analytical data in place

Analytics

Elasticsearch

Index and document ingestion

MongoDB

Collection and document ingestion

Document

Google Drive

Docs, sheets, and files from Drive

Cloud

Hugging Face

Datasets from the Hugging Face Hub

Datasets

pandas

Ingest DataFrames directly

DataFrame

Apache Parquet

Native Parquet file ingestion

Columnar

Apache Arrow

Arrow, Feather, and IPC via ArrowIngestor

Columnar

Streams

Kafka, RabbitMQ, Kinesis, and Pulsar message streams

Kafka · Kinesis

Git Repositories

Ingest code and docs from any Git remote

Repo

Email

Mailbox ingestion over IMAP and POP3

IMAP · POP3

MCP Resources

Ingest resources from any MCP server

MCP

Files & Documents

PDF, DOCX, PPTX, HTML, CSV, JSON, YAML, Excel, XML

10+ formats

JSON Lines

New v0.7.0

Line-delimited JSON (JSONL/NDJSON) file ingestion

NDJSON

Databases

PostgreSQL, MySQL, SQLite, Oracle, SQL Server

SQL

Export & Interchange

Standards-based outputs for regulators, downstream analytics, and other graph systems.

Neo4j Bulk CSV

Bulk CSV export for neo4j-admin import

Cypher import

ArangoDB AQL

ArangoDB export support

AQL

RDF

Turtle, JSON-LD, N-Triples, and RDF/XML

Turtle · JSON-LD

OWL Ontologies

OWL/XML with SHACL shapes and SKOS vocabularies

OWL

Parquet & Arrow

Columnar exports for analytics pipelines

Columnar

CSV · JSON · YAML

Plain interchange formats for everything else

Tabular

Don't see your stack?

Request the connector you need, or bring your engineers to the Discord, where teams from healthcare, finance, legal, and defense are building on a governed context layer.

Request an integration Join Discord Read the docs