Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.agno.com/llms.txt

Use this file to discover all available pages before exploring further.

Document readers extract text from various file formats for ingestion into your knowledge base. Agno auto-detects file types or lets you specify readers explicitly. 
from agno.knowledge.knowledge import Knowledge
from agno.vectordb.pgvector import PgVector

knowledge = Knowledge(
    vector_db=PgVector(table_name="docs", db_url="postgresql://..."),
)

# Auto-detect file type
knowledge.insert(path="./documents/report.pdf")
knowledge.insert(path="./data/records.csv")
knowledge.insert(url="https://example.com/page.html")

​
Supported Readers

FormatReaderUse Case
PDFPDFReaderDocuments, reports, papers
CSVCSVReaderTabular data, spreadsheets
JSONJSONReaderAPI responses, configs
MarkdownMarkdownReaderDocumentation, notes
PowerPointPPTXReaderPresentations
WordDocxReaderWord documents
HTMLHTMLReaderWeb pages
ArxivArxivReaderAcademic papers
YouTubeYouTubeReaderVideo transcripts
FirecrawlFirecrawlReaderWeb scraping
TavilyTavilyReaderWeb research

​
Examples by Format

​
PDF Files

cookbook/07_knowledge/readers/pdf_reader_async.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.pdf import PDFReader
from agno.vectordb.pgvector import PgVector

knowledge = Knowledge(
    vector_db=PgVector(table_name="pdfs", db_url="postgresql://..."),
)

# Add PDF from path
knowledge.insert(
    path="./documents/report.pdf",
    reader=PDFReader(),
)

# Add PDF from URL
knowledge.insert(
    url="https://example.com/whitepaper.pdf",
)

​
Password-Protected PDFs

cookbook/07_knowledge/readers/pdf_reader_password.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.pdf import PDFReader

knowledge = Knowledge(vector_db=...)

knowledge.insert(
    path="./secure.pdf",
    reader=PDFReader(password="secret123"),
)

​
CSV Files

cookbook/07_knowledge/readers/csv_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.csv import CSVReader

knowledge = Knowledge(vector_db=...)

# Basic CSV
knowledge.insert(
    path="./data/customers.csv",
    reader=CSVReader(),
)

​
CSV with Field Labels

Add column names as context.
cookbook/07_knowledge/readers/csv_field_labeled_reader.py
from agno.knowledge.reader.csv import CSVReader

reader = CSVReader(
    include_field_labels=True,
    row_template="Customer {name} from {city} purchased {product}",
)

​
JSON Files

cookbook/07_knowledge/readers/json_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.json import JSONReader

knowledge = Knowledge(vector_db=...)

knowledge.insert(
    path="./data/config.json",
    reader=JSONReader(),
)

​
Markdown Files

cookbook/07_knowledge/readers/markdown_reader_async.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.markdown import MarkdownReader

knowledge = Knowledge(vector_db=...)

# Add all markdown files from directory
knowledge.insert(
    path="./docs/",
    reader=MarkdownReader(),
)

​
PowerPoint Presentations

cookbook/07_knowledge/readers/pptx_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.pptx import PPTXReader

knowledge = Knowledge(vector_db=...)

knowledge.insert(
    path="./presentations/quarterly_review.pptx",
    reader=PPTXReader(),
)

​
Arxiv Papers

Automatically fetch and parse academic papers.
cookbook/07_knowledge/readers/arxiv_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.arxiv import ArxivReader

knowledge = Knowledge(vector_db=...)

# Add paper by Arxiv ID
knowledge.insert(
    topic="2301.00234",  # Arxiv ID
    reader=ArxivReader(),
)

# Search and add papers
knowledge.insert(
    topic="transformer architecture attention",
    reader=ArxivReader(max_results=5),
)

​
Web Content with Firecrawl

cookbook/07_knowledge/readers/firecrawl_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.firecrawl import FirecrawlReader

knowledge = Knowledge(vector_db=...)

knowledge.insert(
    url="https://docs.example.com/",
    reader=FirecrawlReader(
        api_key="your-key",
        crawl_subpages=True,
    ),
)

​
Web Research with Tavily

cookbook/07_knowledge/readers/tavily_reader.py
from agno.knowledge.knowledge import Knowledge
from agno.knowledge.reader.tavily import TavilyReader

knowledge = Knowledge(vector_db=...)

knowledge.insert(
    topic="Latest developments in quantum computing",
    reader=TavilyReader(api_key="your-key"),
)

​
Run Examples

git clone https://github.com/agno-agi/agno.git
cd agno/cookbook/07_knowledge/readers

# PDF
python pdf_reader_async.py

# CSV
python csv_reader.py

# Arxiv
python arxiv_reader.py