Limited-Time OfferGet $1 worth of free queries and upload up to 5 documents for free — no credit card required!Sign up

retrievrSign inProfile

Ask and You Shall Retrieve

One-line RAG for your agent.
No vector database. No embedding pipeline. No indexing.
Paste this line to your agent and ask questions.
Retrievr provides relevant information for your agent to work with.

Fetch https://retrievr.io/install and follow it.

Works with the agent you already have:

Claude Code
Codex
Cursor
pi
opencode
Kiro
Hermes
and others

USAGE

WHAT YOUR AGENT ASKED

WHAT CAME BACK FROM RETRIEVR

The Divine Comedy

Alighieri, Dante

INFERNO > CANTO II. · p. 23

“

YOUR AGENTS ANSWER USING THIS INFORMATION

BENEFITS

Lowered token cost: Your agent gets the passages that answer the question, not the document they came from, so nothing is spent reading what the model will not use. The same answer costs a fraction of the tokens it would have taken to find it.
Document aware RAG: The stack fits the document rather than the other way round: the embedding model and the index, the chunk size, how each file type is read and which language the text is in. EPUB in spine order, PDF page by page, passages packed to the embedder's ceiling, and the model that suits the text instead of one setting forced on all of it.
Reduced hallucination: Every passage arrives with its source, author, section, character range and licence, so the agent quotes what it found instead of reconstructing it from memory. A made-up answer becomes visible instead of plausible.