txtai

RAG Pipeline Example

from txtai import Embeddings
from txtai.pipeline import RAG

# Create embeddings database
embeddings = Embeddings(path="sentence-transformers/all-MiniLM-L6-v2")

# Create RAG pipeline
rag = RAG(
    embeddings,
    "microsoft/DialoGPT-medium",  # LLM for generation
    citation=True  # Enable citation support
)

# Add documents
documents = [
    "Python is a high-level programming language...",
    "Machine learning is a subset of AI...",
    "Vector databases store embeddings..."
]
embeddings.index(documents)

# Run RAG query
result = rag("What is Python?")
print(result)
                

txtai RAG pipelines automatically handle context retrieval, LLM prompting, and response generation.

Agent Example

from datetime import datetime
from txtai import Agent

# Define tools
def today() -> str:
    """Gets current date and time"""
    return datetime.today().isoformat()

wikipedia = {
    "name": "wikipedia",
    "description": "Searches a Wikipedia database",
    "provider": "huggingface-hub",
    "container": "neuml/txtai-wikipedia"
}

# Create agent
agent = Agent(
    model="Qwen/Qwen3-4B-Instruct-2507",
    tools=[today, wikipedia, "websearch"],
    max_steps=10
)

# Use agent
result = agent("Which city has highest population, Boston or New York?")
                

Agents excel at complex tasks requiring multiple tools and reasoning steps.

Quick Start

# Install txtai
pip install txtai

# Install with all dependencies
pip install txtai[pipeline]

# Basic usage
from txtai import Embeddings

# Create embeddings
embeddings = Embeddings()

# Index documents
documents = [
    "US tops 5 million confirmed virus cases",
    "Canada's last fully intact ice shelf has collapsed",
    "Beijing mobilises invasion craft along coast"
]

embeddings.index(documents)

# Search
results = embeddings.search("climate change", limit=2)
print(results)