- Graphwise Platform Documentation
- GraphRAG
- GraphRAG Configuration Guide
- Graphwise Graph Modeling Integration
Graphwise Graph Modeling Integration
01/10/2026
GraphRAG version 1.3
Graphwise Graph Modeling (formerly PoolParty) supplies the taxonomy management engine, enterprise custom schema metadata, and semantic core backbone properties for the pipeline execution loop.
To establish a secure line of communication with the modeling platform, open the n8n workspace dashboard and locate the centralized Configuration workflow. Assign the following connectivity attributes inside the JavaScript configuration code block:
graphModelingServerUrlThe base network access endpoint of your Graph Modeling platform instance.
graphModelingProjectIdThe target validation UUID matching your specific taxonomy project.
// Graph Modeling configurationgraphModelingServerUrl: "https://poolparty-dev.example.com",graphModelingProjectId: "00000000-0000-0000-0000-000000000000",
Authentication details must be managed safely outside the raw workflow logic:
Navigate to the Credentials Sidebar window inside the n8n application dashboard.
Populate the Graph Modeling API record and define your preferred authentication protocol (Basic Auth or OAuth 2.0 validation tokens).
Map these created credentials directly to the following workflow orchestration elements to authorize automated calls:
Graph Modeling Concept Enricher - Sync
Graph Modeling Expansion - Async
Fallback Language Translation
By default, the query parsing framework translates conceptual components into English (EN). To change this fallback language to your preferred corporate ISO localization string (e.g., DE, FR, ES), follow these steps:
Open the Main workflow canvas.
Locate the Call 'Graph Modeling concept enricher' -> Answer Concept Details node element.
Update the language identification parameter to your target string.
Contextual Sensitivity Limits
To fine-tune extraction thresholds and balance contextual query sensitivity, modify the data object property values directly inside the Step - Graph Modeling Concept Enricher - Sync sub-workflow configuration node.
{
"numberOfConcepts": 10,
"numberOfTerms": 20,
"language": "en",
"corpusScoring": "TFIDF",
"frequencyThreshold": 2
} The graph extraction framework pulls corresponding definitions dynamically based on the configured properties string array.
"conceptProperties": [ "skos:prefLabel", "skos:altLabel", "skos:hiddenLabel", "skos:definition", "skos:scopeNote", "skos:example", "skos:broader", "skos:narrower", "skos:related" ]
Note
Custom proprietary relationship schema URIs (e.g., https://example.com/MyScheme/MyRelation) can be appended directly into this list.
To be able to use any non-SKOS property of concepts in the LLM context, provide a description in the form of a single short sentence for the LLM to understand how the value relates to the original concept for each of the relations/attributes ypu want to add, e.g.:
"conceptPropertyLlmContext": {
"https://example.com/MyScheme/MyRelation": "This is the description of the custom relation for the LLM",
"https://example.com/MyScheme/MyAttribute": "This is the description of the custom attribute for the LLM"
}