Semantica plugin
Semantica plugin: 17 skills and 3 agents for knowledge graphs, reasoning and data provenance in Claude Code, Cursor and Codex.
- 14,448 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Tools read connected data sources and write to the graph; queries may go to model providers.
Install
- Works in
Manual install
pip install semanticaThe first README command installs the library. The Claude Code plugin is loaded from a clone with `claude --plugin-dir ./plugins`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Tools read connected data sources and write to the graph; queries may go to model providers.
This is third-party code. Review the repository files before installing.
What it does
Semantica, a Python library for knowledge graphs and AI context, ships a plugin set for agent editors. The `plugins/` folder holds 17 skills (`extract`, `ingest`, `query`, `ontology`, `reason`, `decision`, `provenance`, `export` and more), three agents (`decision-advisor`, `explainability`, `kg-assistant`) and hooks. Manifests exist for Claude Code, Cursor, Codex, Cline, Windsurf, Continue, OpenClaw, VS Code and pi. A separate MCP server runs as `semantica-mcp` or `python -m semantica.mcp_server`. It adds nodes and relations, finds decision precedents, runs rule sets and exports the graph to RDF, JSON and Parquet.
Who it is for. Data engineers and AI developers building a knowledge graph who want to run it from an agent instead of hand-written Python.
Good fit when
- Entities and relations are pulled from contracts and reports into a graph.
- An agent needs a decision trail: who chose an option, when and on what basis.
- A team checks where a claim in the graph came from.
- The graph is exported to OWL, Cypher or JSON-LD for another system.
Not a fit when
- A plain vector store is enough, with no graph or ontologies.
- Python 3.11 or newer is unavailable on the machine.
- The data cannot go to local stores and third-party model providers.
Example request
Extract entities and relations from `contract.pdf`, build a graph and show the provenance chain for the penalty amount.Limitations
The plugins run on top of the installed library: Python 3.11 and `pip install semantica` are required, and for Claude Code the plugin is loaded from a repository clone with `claude --plugin-dir ./plugins`. The project is large and some features come as extras such as `semantica[documents]`. The `delete_node` MCP tool archives a node rather than erasing it. Plugin descriptions follow version 0.4.0, so compare against the repository's `plugins/` folder.
How to disable. Drop `--plugin-dir ./plugins` from the Claude Code launch or remove the plugin from another editor's settings, then delete the `semantica` MCP entry. `pip uninstall semantica` removes the library.
FAQ
What does the plugin include?
17 skills, 3 agents (`decision-advisor`, `explainability`, `kg-assistant`) and hooks in the `plugins/` folder.
How do I connect it to Claude Code?
Install the library with `pip install semantica`, clone the repository and start `claude --plugin-dir ./plugins`.
Which editors are supported?
Separate manifests exist for Claude Code, Cursor, Codex, Cline, Windsurf, Continue, OpenClaw, VS Code and pi.
How do I start the MCP server?
Run `semantica-mcp` or `python -m semantica.mcp_server`, then add it to the client's `mcpServers`.
Which tools does the MCP server expose?
Among them `extract_entities`, `add_entity`, `query_graph`, `run_reasoning`, `find_precedents` and `export_graph`, 15 in total.
Discussion
No questions or reviews yet.
Related
PluginsA plugin with two skills over a semantic metrics layer: the agent writes models in YAML and queries them through SQL.
A Claude Code and Codex plugin that builds an ontology over your tables and files and refines it from the agent's work history
A knowledge graph with an MCP server: the agent searches concepts by meaning and sees how well each is supported.
Official Pinecone plugin: an agent creates indexes, searches them, builds a RAG assistant and cites your files.
A CLI and MCP server that extracts structured knowledge from documents: graphs, hypergraphs, and fact search with sources
Around 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Semantica (this card) | no | 14,448 | ||
| Cognee | no | 33,093 | agent memory on a graph; Semantica adds ontologies, reasoning rules and provenance chains | |
| Graphiti MCP | no | API key | 32,574 | a temporal graph for agent memory; here 17 skills and plugins for nine editors |
README in short
The README lists native plugins for major editors, an MCP server with 15 tools, a REST API and a CLI with command groups such as `ingest`, `reason`, `provenance` and `export`.