Configure your LLM
Connect any LLM — OpenAI, Anthropic, Groq, OpenRouter, or a custom endpoint. The lab uses your API key to run two parallel calls: one without any context (LLM alone) and one enriched with your knowledge graph (LLM + Khognitive).
Build your personal knowledge graph
This is the knowledge Khognitive will inject into the LLM during Step 3. The more specific and real your context, the more dramatic the difference you will see in the comparison. Pick any of the three modes below — you can combine them freely and add more at any point during the session.
📝 Free Text
✏️ Manual Facts
🎬 Demo Scenarios
Paste any text you own — company facts, internal documents, meeting notes, product briefs, research summaries, or any domain knowledge relevant to what you want to test. Click Extract Knowledge and Khognitive will automatically identify the relationships inside your text. You will see a table of extracted triples you can review, edit, or delete before adding them to the graph. Nothing is ingested without your approval.
Extracted Triples — review & edit
| Subject | Relation | Object | Conf | |
|---|---|---|---|---|
triples will be added to your knowledge graph
Type facts one by one — use this when you already know exactly what relationships you want to test. Each row is a triple: a subject, a relation, and an object. Example: SARAH_CHEN → LEADS → ACME_CORP. Keep subjects and objects as short, specific nouns. Keep relations as active verbs (LEADS, OWNS, TARGETS, RAISED, REQUIRES). Add as many rows as you need, then click Add to Graph.
→
Don't have your own data ready? Start with one of these pre-built scenarios to see Khognitive in action immediately. Click Load on any scenario to add its knowledge to your graph, then go to Step 3 and ask questions about it. You can load multiple scenarios and mix them with your own context from the Free Text or Manual Facts tabs — everything accumulates in the same session graph.
Your Knowledge Graph
Nodes
Edges
Rel.
Add context to see graph grow
Ingest Log
Graph: ·
·
— go to Step 2 to load context first
LLM Alone
Querying…
Run a test to see how the LLM responds with no external knowledge.
LLM + Khognitive
Querying with knowledge graph…
Run a test to see how Khognitive enriches the response with your graph.
Knowledge Retrieved by Khognitive
nodes
edges
rel. types
lines injected
No knowledge lines retrieved. Try loading context in Step 2 first.
Run at least one test in Step 3 to see live analytics.
Total Runs
Avg Token Savings
Avg Latency Delta
Total Facts Injected
📈 Knowledge Graph Growth
🔢 Token Consumption per Run
⏱ Latency Timeline
All Runs ()
| # | Question | Alone Tokens | +Khog Tokens | Alone ms | +Khog ms | Facts |
|---|---|---|---|---|---|---|