Blink RAG

Search a Blink project's knowledge base using semantic search or AI-synthesized answers. Upload documents to build a searchable knowledge base. Use when you need to answer questions from a corpus of documents, FAQs, or any structured knowledge. Requires BLINK_PROJECT_ID secret.

blink-new 5c63d10 1.5 KB Updated

File contents

Blink RAG — Knowledge Base

Semantic search over your Blink project's document collection using blink rag. Upload documents once, search them forever. AI-enhanced search synthesizes an answer from relevant chunks.

Setup (one-time)

blink secrets set BLINK_PROJECT_ID proj_xxx

Get your project ID from blink.new → your project → Settings.

Search the knowledge base

blink rag search $BLINK_PROJECT_ID "how does the refund policy work"

AI-synthesized answer (recommended for Q&A)

blink rag search $BLINK_PROJECT_ID "what are the pricing tiers" --ai

Search with more results

blink rag search $BLINK_PROJECT_ID "billing and payments" --ai --limit 10

Upload a document

blink rag upload $BLINK_PROJECT_ID ./docs/faq.md

Upload to a specific collection

blink rag upload $BLINK_PROJECT_ID ./docs/refund-policy.pdf --collection coll_xxx

List collections

blink rag collections $BLINK_PROJECT_ID

Command signatures

blink rag search <project_id> <query> [--ai] [--limit N]
blink rag upload <project_id> <file> [--collection <collection_id>]
blink rag collections <project_id>

blink-new/openclaw/tree/main/skills/blink-rag commit 5c63d10422

Frequently asked questions

npx skillmds@latest add blink-new/blink-rag