GraphRAG · Scale Edition
Financial Intelligence
100M+ Token Source Data → 692 Node Knowledge Graph: LLM-Only vs Basic RAG vs TigerGraph GraphRAG
TigerGraph Cloud (Savanna)
⚙ TigerGraph Cloud Config
P1
LLM-Only
No retrieval · baseline
IDLE
Response will appear here…
Tokens
Latency
Cost/query
Groundedness F1
LLM Judge
Retrieval
None
P2
Basic RAG
Vector embeddings · baseline
IDLE
Response will appear here…
Tokens
Latency
Cost/query
Groundedness F1
LLM Judge
Retrieval
Vector
P3
GraphRAG
TigerGraph Cloud · multi-hop
IDLE
Response will appear here…
Tokens
Latency
Cost/query
Groundedness F1
LLM Judge
Retrieval
Graph

⚡ TigerGraph Cloud Setup (from your guide)
1
Create account at tgcloud.io via email or GitHub, verify, log into dashboard.
2
Create a cluster — click "Create Cluster", choose Free Tier, pick your region, name it.
3
Open GraphStudio — Tools → GraphStudio. Design your schema (vertices + edges).
4
Load data via CSV upload + Loading Jobs, or REST / pyTigerGraph.
5
Write GSQL query, install it — this generates a REST endpoint automatically.
6
Generate RESTPP token in Admin Portal → User Management → paste it above.
7
Endpoint format: https://<cluster>.i.tgcloud.io/query/<GRAPH>/<QUERY>
8
Test in Postman with GET + Authorization: Bearer YOUR_TOKEN before wiring here.