Haystack: Pipelines, Components and RAG Reference Guide

# Core + the integrations you'll need
pip install haystack-ai                          # the v2.x package (do NOT install old "farm-haystack")
pip install chroma-haystack                      # vector store
pip install -q "sentence-transformers"           # local embedder

# Provider-specific packages
pip install openai-haystack
pip install anthropic-haystack
pip install cohere-haystack

# env
export OPENAI_API_KEY=sk-...
export ANTHROPIC_API_KEY=sk-...
export HAYSTACK_TELEMETRY_ENABLED=false          # opt out of usage telemetry

# Run a saved pipeline as a service (Hayhooks)
pip install hayhooks
hayhooks pipeline deploy ./pipeline.yaml

from pathlib import Path
from haystack import Pipeline
from haystack.components.converters import TextFileToDocument, PyPDFToDocument
from haystack.components.preprocessors import DocumentCleaner, DocumentSplitter
from haystack.components.writers import DocumentWriter
from haystack.components.embedders import SentenceTransformersDocumentEmbedder
from haystack_integrations.document_stores.chroma import ChromaDocumentStore

store = ChromaDocumentStore(persist_path="./chroma")

idx = Pipeline()
idx.add_component("txt",    TextFileToDocument())
idx.add_component("clean",  DocumentCleaner())
idx.add_component("split",  DocumentSplitter(split_by="word", split_length=180, split_overlap=30))
idx.add_component("embed",  SentenceTransformersDocumentEmbedder(model="BAAI/bge-small-en-v1.5"))
idx.add_component("write",  DocumentWriter(document_store=store))

idx.connect("txt.documents",   "clean.documents")
idx.connect("clean.documents", "split.documents")
idx.connect("split.documents", "embed.documents")
idx.connect("embed.documents", "write.documents")

idx.run({"txt": {"sources": list(Path("./data").glob("*.txt"))}})

from haystack import Pipeline
from haystack.components.embedders import SentenceTransformersTextEmbedder
from haystack.components.builders import PromptBuilder
from haystack.components.generators import OpenAIGenerator
from haystack_integrations.components.retrievers.chroma import ChromaEmbeddingRetriever

TEMPLATE = """
Answer using only the context. If unsure, say so.

Context:
{% for d in documents %}- {{ d.content }}
{% endfor %}

Question: {{ question }}
Answer:""".strip()

rag = Pipeline()
rag.add_component("embed",   SentenceTransformersTextEmbedder(model="BAAI/bge-small-en-v1.5"))
rag.add_component("retrieve", ChromaEmbeddingRetriever(document_store=store, top_k=4))
rag.add_component("prompt",  PromptBuilder(template=TEMPLATE))
rag.add_component("llm",     OpenAIGenerator(model="gpt-4o-mini"))

rag.connect("embed.embedding",      "retrieve.query_embedding")
rag.connect("retrieve.documents",   "prompt.documents")
rag.connect("prompt.prompt",        "llm.prompt")

out = rag.run({
    "embed":  {"text": "What is the WFH policy?"},
    "prompt": {"question": "What is the WFH policy?"},
})
print(out["llm"]["replies"][0])

from haystack.components.agents import Agent
from haystack.components.generators.chat import OpenAIChatGenerator
from haystack.tools import Tool
from haystack.dataclasses import ChatMessage

def get_weather(city: str) -> str:
    """Return current weather for a city."""
    return f"{city}: 22 C, clear."

weather = Tool.from_function(
    function=get_weather,
    name="get_weather",
    description="Look up current weather for a city.",
)

agent = Agent(
    chat_generator=OpenAIChatGenerator(model="gpt-4o-mini"),
    tools=[weather],
    system_prompt="You are a precise assistant. Use tools when helpful.",
    exit_conditions=["text"],            # stop on first plain-text reply
    max_agent_steps=8,
)

result = agent.run(messages=[ChatMessage.from_user("What's the weather in Berlin?")])
print(result["messages"][-1].text)

from pathlib import Path
from haystack import Pipeline
from haystack.components.converters import TextFileToDocument
from haystack.components.preprocessors import DocumentSplitter
from haystack.components.embedders import (
    SentenceTransformersDocumentEmbedder, SentenceTransformersTextEmbedder,
)
from haystack.components.writers import DocumentWriter
from haystack.components.retrievers.in_memory import InMemoryEmbeddingRetriever
from haystack.components.builders import PromptBuilder
from haystack.components.generators import OpenAIGenerator
from haystack.document_stores.in_memory import InMemoryDocumentStore

store = InMemoryDocumentStore()

# Index
idx = Pipeline()
idx.add_component("read",  TextFileToDocument())
idx.add_component("split", DocumentSplitter(split_by="word", split_length=180))
idx.add_component("embed", SentenceTransformersDocumentEmbedder(model="BAAI/bge-small-en-v1.5"))
idx.add_component("write", DocumentWriter(document_store=store))
idx.connect("read.documents", "split.documents")
idx.connect("split.documents", "embed.documents")
idx.connect("embed.documents", "write.documents")
idx.run({"read": {"sources": list(Path("data").glob("*.txt"))}})

# Query
rag = Pipeline()
rag.add_component("q_emb",   SentenceTransformersTextEmbedder(model="BAAI/bge-small-en-v1.5"))
rag.add_component("retrieve", InMemoryEmbeddingRetriever(document_store=store, top_k=4))
rag.add_component("prompt",  PromptBuilder(template="Q: {{q}}\nDocs: {{docs}}\nA:"))
rag.add_component("llm",     OpenAIGenerator(model="gpt-4o-mini"))
rag.connect("q_emb.embedding", "retrieve.query_embedding")
rag.connect("retrieve.documents", "prompt.docs")
rag.connect("prompt.prompt", "llm.prompt")

print(rag.run({"q_emb": {"text": "Summarise the handbook."},
               "prompt": {"q": "Summarise the handbook."}})["llm"]["replies"][0])