Features · Connect · RAG

Answers from your documents, with citations.

Name: SHIPIT Agent
Author: SHIPIT

Wire shipit's RAG into an agent with one parameter, rag=. The agent gains rag_search, rag_fetch_chunk and rag_list_sources, and its prompt asks it to cite the passages it used.

RAG with an agent Standalone RAG

rag.pylive

from shipit_agent import Agentfrom shipit_agent.rag import RAG, HashingEmbedderrag = RAG.default(embedder=HashingEmbedder(dimension=512))rag.index_text("Shipit supports Python 3.10+.", source="readme.md")rag.index_text("Agents stream events in real time.", source="streaming.md")agent = Agent(llm=my_llm, rag=rag)

$ pythonexample output

running…

How it works

Index

text, files, PDFs

Hybrid search

vectors plus BM25

Agent

searches and reads

Answer

with [N] citations

See it run

The code is a checked example from the docs; the output is an example of what it prints.

rag.py

from shipit_agent import Agent
from shipit_agent.rag import RAG, HashingEmbedder

rag = RAG.default(embedder=HashingEmbedder(dimension=512))
rag.index_text("Shipit supports Python 3.10+.", source="readme.md")
rag.index_text("Agents stream events in real time.", source="streaming.md")

agent = Agent(llm=my_llm, rag=rag)
result = agent.run("What Python version does Shipit support?")

print(result.output)

$ pythonexample output

running…

Go deeper

6 more ways to use it

Each example is from the docs and checked against the library.

Quickstart

Read in the docs

example.py

from shipit_agent import Agent
from shipit_agent.rag import RAG, HashingEmbedder

rag = RAG.default(embedder=HashingEmbedder(dimension=512))
rag.index_text("Shipit supports Python 3.10+.", source="readme.md")
rag.index_text("Agents stream events in real time.", source="streaming.md")

agent = Agent(llm=my_llm, rag=rag)
result = agent.run("What Python version does Shipit support?")

print(result.output)
# "Shipit supports Python 3.10+. [1]"

for src in result.rag_sources:
    print(f"[{src.index}] {src.source} (chunk {src.chunk_id}, score {src.score:.2f})")
    print(f"    {src.text}")

What you get

One parameter
rag= adds the search tools and the citation instructions.
Hybrid retrieval
Vector and keyword search, fused; a reranker is optional.
Your backends
Swap in Chroma, Qdrant, pgvector or your own store.
Shared indexes
One RAG can serve many agents at once.

Agent parameters

Parameter	Default	What it does
rag	None	A RAG the agent searches and cites. RAG with an agent.

More in Connect

Next: Sessions and memory

from shipit_agent import Agentfrom shipit_agent.rag import RAG, HashingEmbedderrag = RAG.default(embedder=HashingEmbedder(dimension=512))rag.index_text("Shipit supports Python 3.10+.", source="readme.md")rag.index_text("Agents stream events in real time.", source="streaming.md")agent = Agent(llm=my_llm, rag=rag)

See it run

The code is a checked example from the docs; the output is an example of what it prints.

rag.py

from shipit_agent import Agent
from shipit_agent.rag import RAG, HashingEmbedder

rag = RAG.default(embedder=HashingEmbedder(dimension=512))
rag.index_text("Shipit supports Python 3.10+.", source="readme.md")
rag.index_text("Agents stream events in real time.", source="streaming.md")

agent = Agent(llm=my_llm, rag=rag)
result = agent.run("What Python version does Shipit support?")

print(result.output)

$ pythonexample output

running…

Go deeper

6 more ways to use it

Each example is from the docs and checked against the library.

Quickstart

Read in the docs

example.py

from shipit_agent import Agent
from shipit_agent.rag import RAG, HashingEmbedder

rag = RAG.default(embedder=HashingEmbedder(dimension=512))
rag.index_text("Shipit supports Python 3.10+.", source="readme.md")
rag.index_text("Agents stream events in real time.", source="streaming.md")

agent = Agent(llm=my_llm, rag=rag)
result = agent.run("What Python version does Shipit support?")

print(result.output)
# "Shipit supports Python 3.10+. [1]"

for src in result.rag_sources:
    print(f"[{src.index}] {src.source} (chunk {src.chunk_id}, score {src.score:.2f})")
    print(f"    {src.text}")

Parameter

Default

What it does

rag

None

A RAG the agent searches and cites. RAG with an agent.