Usage — API / CLI / MCP¶
jena-mcp exposes the same capability three ways: as MCP tools an agent calls,
as a Python API (JenaApi) the operator import, and as a CLI. The complete tool
surface is summarized in Overview.
As an MCP server¶
Once deployed, the server registers one typed tool per operation that cover the SPARQL Protocol, the Graph Store Protocol, and Fuseki administration:
| Tool | Purpose |
|---|---|
jena_query / jena_update |
SPARQL SELECT/ASK/CONSTRUCT/DESCRIBE and INSERT/DELETE/LOAD/CLEAR |
jena_get_graph / jena_put_graph / jena_post_graph / jena_delete_graph |
Graph Store Protocol: read, replace, merge, delete |
jena_ping, jena_server_info, jena_stats, jena_metrics, jena_list_datasets, jena_dataset_info, jena_create_dataset, jena_delete_dataset, jena_set_dataset_state, jena_list_tasks, jena_task_info, jena_backup, jena_compact |
server administration |
Example agent prompts that map onto these tools:
- "List the datasets on the Fuseki server" →
jena_list_datasets - "Run this SELECT query against dataset
ds" →jena_query - "Load this Turtle into the default graph of
ds" →jena_post_graph
As a Python API¶
JenaApi (exported as Api) is a requests-based facade over a Fuseki endpoint,
organized by protocol.
from jena_mcp.api_client import Api
api = Api(
base_url="https://jena.example.invalid",
username="admin",
password="admin",
)
# SPARQL reads
results = api.query("ds", "SELECT * WHERE { ?s ?p ?o } LIMIT 10")
# Graph Store Protocol read (default graph as Turtle)
turtle = api.get_graph("ds")
# Server administration (read-only)
datasets = api.list_datasets()
info = api.server_info()
stats = api.stats("ds")
Build a client straight from the environment:
Writes¶
SPARQL updates and Graph Store writes mutate the dataset:
api.update("ds", "INSERT DATA { <urn:a> <urn:knows> <urn:b> }")
api.put_graph("ds", "<urn:a> <urn:p> <urn:o> .", graph="urn:g") # replace
api.post_graph("ds", "<urn:a> <urn:p> <urn:o2> .", graph="urn:g") # merge
# Administration writes
api.create_dataset("ds2", db_type="tdb2")
api.backup("ds")
api.compact("ds", delete_old=True)
As a CLI¶
Run the MCP server directly with the jena-mcp console script (see
Deployment for transports), or invoke the package as a module:
configuration.