CKG — NVIDIA — AI
NVIDIA AI
Developer Docs.
Deterministic traversal context for your agent.
20 NVIDIA AI Developer Documentation Domains. 998 nodes. Your agent traverses declared relationships across the full NVIDIA AI stack — not text it infers from.
20 domains 998 nodes read-only
How it's built
Every edge is a decision.
REQUIRES
Hard prerequisite — cannot function without it. Drives agent sequencing.
ENABLES
Unlocks a capability. Not strictly required — optimization paths.
RELATES_TO
Conceptual proximity. Not a dependency. Use for disambiguation.
IMPLEMENTS
Concrete instantiation of an abstraction. Maps architecture.
Confidence defers to null — not wrong, just unreviewed. This is the scaffold your agent runs before you add your domain layer on top.
KRB Benchmark  ·  v0.6.2  ·  open & reproducible
The eval for structured knowledge retrieval.
# System Macro F1 Tokens/q 5-Hop F1
1 CKG (ckg-mcp v0.7.6)
0.488
252 0.786
2 RAG (text-embedding-3-small)
0.123
2,982 0.170
3 GraphRAG (MS global mode)
0.120
3,450 —
CKG F1 improves with hop depth — 0.37 → 0.77 from hop 0 to hop 5. RAG stays flat at ~0.13 regardless of depth. Retrieval has no mechanism for traversing a chain.
↗ danyarm/krb-leaderboard — open & reproducible
Token efficiency
11× fewer tokens.
Context you save, you keep.
269 CKG · tokens/query
vs
2,982 RAG · tokens/query
A grounding pass that costs 269 tokens leaves your context window open for reasoning. Every token saved is a token you can spend on what actually matters to your agent.
Context as a Service
Your domain, compressed.
This is a rapid-start scaffold — not a be-all-end-all. Layer your domain on ours, or let us build a CKG tuned to your exact knowledge gaps.
Layer 1
Context optimization
Free. 97 domains. Install and go.
Layer 2
Agent grounding
Custom domain CKG built for your stack.
Layer 3
Sealed appliance
Bundled CKG + query server for your team.
Start a conversation → graphifymd.com
1 / 5
CKG — NVIDIA AI  ·  New in v0.5.0
Small Model.
Big Brain.
THE SHIFT
WITHOUT CKG
70B
High cost  ·  hallucination risk
CKG + QWEN 2.5 14B · OLLAMA
14B
Local  ·  sovereign  ·  grounded
ASK_NVIDIA()  ·  V0.5.0
Natural language in. Graph-grounded answer out. Model never leaves the machine.
Tokens / query 269 vs 2,982
Cost / 1K queries $7.81 vs $76.23
F1 accuracy 0.471 vs 0.123
Inference Local Ollama · $0 API