Skip to main content

​
vrin insert

Insert text content into the knowledge base. VRIN automatically chunks the text, extracts structured facts, and indexes everything in the knowledge graph. 
vrin insert "ACME Corp reported $50M revenue in Q4 2025, up 23% YoY." --title "ACME Q4"

​
Options

FlagShortDefaultDescription
--title-t"Untitled"Document title
--tagTags (repeatable: --tag finance --tag earnings)
--metadataMetadata as JSON string
--wait / --no-wait--waitWait for processing to complete
--dry-runfalseShow payload without sending
--jsonForce JSON output

​
Async mode

By default, insert waits for fact extraction to complete. Use --no-wait to get a job ID immediately: 
vrin insert "long document content..." --title "Report" --no-wait --json
# {"ok": true, "data": {"job_id": "job_abc123"}}

# Check status later
vrin job job_abc123

​
Tags and metadata

vrin insert "content" \
  --title "Q4 Earnings" \
  --tag finance \
  --tag quarterly \
  --metadata '{"source": "SEC filing", "period": "Q4 2025"}'

​
vrin upload

Upload a file to the knowledge base. Supports PDF, CSV, TXT, MD, and DOCX. 
vrin upload ./earnings_report.pdf

​
Options

FlagDefaultDescription
--save-to-memory / --no-save-to-memory--save-to-memoryPersist extracted knowledge permanently
--wait / --no-wait--waitWait for processing to complete
--jsonForce JSON output

​
Examples

# Upload and wait for processing
vrin upload ./report.pdf

# Upload without waiting
vrin upload ./report.pdf --no-wait --json
# {"ok": true, "data": {"job_id": "job_xyz789"}}

# Upload for temporary analysis (not persisted)
vrin upload ./draft.pdf --no-save-to-memory

​
vrin bulk-insert

Insert multiple items from a JSON file. Uses adaptive concurrency for fast batch processing. 
vrin bulk-insert ./papers.json

​
Input format

The JSON file must contain an array of objects: 
[
  {
    "content": "First document text...",
    "title": "Document 1",
    "tags": ["research"],
    "metadata": {"author": "Smith"}
  },
  {
    "content": "Second document text...",
    "title": "Document 2"
  }
]
Only content is required. title, tags, and metadata are optional.

​
Options

FlagDefaultDescription
--concurrency8Initial concurrent workers
--max-concurrency25Maximum concurrent workers (auto-scales)
--tagDefault tags applied to all items
--jsonForce JSON output

​
Example

# Insert 100 research papers with progress output
vrin bulk-insert ./papers.json --tag research --tag 2025

# Output:
# [1/100] completed: Adaptive-RAG: Learning to Adapt...
# [2/100] completed: Self-RAG: Learning to Retrieve...
# ...

​
vrin job

Check the status of an async job (from --no-wait inserts or uploads). 
vrin job job_abc123

​
Options

FlagDescription
--waitPoll until the job completes
--jsonForce JSON output

​
Example

# Check once
vrin job job_abc123 --json
# {"ok": true, "data": {"status": "processing", "progress": 0.6}}

# Wait for completion
vrin job job_abc123 --wait --json
# {"ok": true, "data": {"status": "completed", "facts_extracted": 47}}