Install opens the AgentOne desktop app. You need the latest version.
Connect AgentOne to LlamaParse to read, search, and pull structured data from your documents. LlamaParse balances cost and accuracy, reading documents at 4x lower token costs and up to 5x higher accuracy.
LlamaParse is especially useful for documents that include:
Complex page layouts with multiple columns or nested sections that need dedicated layout parsing, not just text extraction
Embedded visuals like charts, images, and tables that require structured visual parsing to extract accurate numbers
Low-quality scans or handwriting that need specialized document processing beyond what general-purpose models provide
LlamaParse offers several state-of-the-art document understanding capabilities. With it, AgentOne can:
Parse entire documents into structured markdown, JSON, or HTML, ready for downstream agentic reasoning
Extract defined schemas of information from long documents to pull out only what is relevant
Index documents and search across them with filesystem-style tools: locate relevant files, grep for pattern matches, read a file in full, or run hybrid retrieval
Classify documents into known types
Split document pages into individual pre-defined sections
LiteParse, an open-source parser from LlamaIndex, works well for text-heavy PDFs with straightforward single-column layouts and no tables, charts, or scanned pages to interpret.
Requires a LlamaParse account.
Tools
addFilesToDirectory
classifyFile
createDirectory
createIndex
estimateFileComplexity
extractFile
findFilesInIndex
generateExtractionConfig
getIndexStatus
getUploadUrl
getUserProjects
grepFileFromIndex
listDirectories
listDirectory
listIndexes
parseFile
parseWithLiteParse
readFileFromIndex
retrieveFromIndex
splitFile
syncIndex
uploadFileByUrl
Only use extensions from developers you trust. AgentOne doesn't control which tools developers make available and can't verify that they'll work as intended or won't change.