Skip to main content
Semantic chunking is a method of splitting documents into smaller chunks by analyzing semantic similarity between text segments using embeddings. It uses the chonkie library to identify natural breakpoints where the semantic meaning changes significantly, based on a configurable similarity threshold. This helps preserve context and meaning better than fixed-size chunking by ensuring semantically related content stays together in the same chunk, while splitting occurs at meaningful topic transitions.
1

Create a Python file

touch semantic_chunking.py
2

Add the following code to your Python file

semantic_chunking.py
import asyncio
from agno.agent import Agent
from agno.knowledge.chunking.semantic import SemanticChunking
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.pdf_reader import PDFReader
from agno.vectordb.pgvector import PgVector

db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"

knowledge = Knowledge(
    vector_db=PgVector(table_name="recipes_semantic_chunking", db_url=db_url),
)
asyncio.run(knowledge.add_content_async(
    url="https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf",
    reader=PDFReader(
        name="Semantic Chunking Reader",
        chunking_strategy=SemanticChunking(similarity_threshold=0.5),
    ),
))

agent = Agent(
    knowledge=knowledge,
    search_knowledge=True,
)

agent.print_response("How to make Thai curry?", markdown=True)
3

Create a virtual environment

Open the Terminal and create a python virtual environment.
python3 -m venv .venv
source .venv/bin/activate
4

Install libraries

pip install -U sqlalchemy psycopg pgvector agno chonkie
5

Run PgVector

docker run -d \
  -e POSTGRES_DB=ai \
  -e POSTGRES_USER=ai \
  -e POSTGRES_PASSWORD=ai \
  -e PGDATA=/var/lib/postgresql/data/pgdata \
  -v pgvolume:/var/lib/postgresql/data \
  -p 5532:5432 \
  --name pgvector \
  agno/pgvector:16
6

Run the script

python semantic_chunking.py

​
Semantic Chunking Params

ParameterTypeDefaultDescription
embedderEmbedderOpenAIEmbedderThe embedder to use for semantic chunking.
chunk_sizeint5000The maximum size of each chunk.
similarity_thresholdfloat0.5The similarity threshold for determining chunk boundaries.