Skip to main content

​
Install

pip install vrin
Requires Python 3.9+.

​
Get an API key

Sign up at vrin.cloud to get your API key, then set it: 
export VRIN_API_KEY=vrin_your_api_key
Or create one from the CLI: 
vrin auth login you@example.com
vrin auth create-key --name "my-project"

​
Insert knowledge

vrin insert "ACME Corp reported $50M revenue in Q4 2025, up 23% YoY. CEO Jane Smith attributed growth to enterprise." --title "ACME Q4"
VRIN automatically chunks the text, extracts structured facts, and indexes everything in the knowledge graph.

​
Query

vrin query "What is ACME's revenue?"

​
Stream responses

vrin query "Summarize ACME's Q4 performance" --stream

for token in client.query("Summarize ACME's Q4 performance", stream=True):
    print(token, end="", flush=True)

​
Upload files

vrin upload ./earnings_report.pdf

client.upload_file("earnings_report.pdf")

​
Bulk insert

Insert many documents at once from a JSON file: 
vrin bulk-insert ./papers.json --tag research

​
Multi-turn conversations

vrin query "What was ACME's Q4 revenue?" --json
# Note the session_id in the response

vrin query "How does that compare to Q3?" --session-id "sess_abc123"

client.start_conversation()
r1 = client.continue_conversation("What was ACME's Q4 revenue?")
r2 = client.continue_conversation("How does that compare to Q3?")
client.end_conversation()

​
For AI agents

VRIN is designed to be called by AI agents. When stdout is not a TTY (piped or called programmatically), output is automatically JSON. Use vrin --describe to get a machine-readable schema of all commands. 
# Auto-discovery for agents
vrin --describe

# Programmatic usage (always JSON)
vrin query "What is ACME's revenue?" --json
vrin facts "ACME" --json

​
Next steps