Skip to main content

Knowhere API

Knowhere API is a Chunking-as-a-Service platform that transforms complex, unstructured documents into structured, RAG-ready data. Our API extracts and preserves the semantic structure of your documents—including paragraphs, tables, headings, and images—delivering high-quality chunks that significantly improve your RAG application's performance.

Core Value Proposition​

Extract structure, not just text.

Traditional document parsing tools output plain text streams, losing critical semantic information. Knowhere API preserves the document's logical hierarchy, relationships between content blocks, and multi-modal elements (tables, images), enabling superior chunking and embedding quality.

Key Features​

Intelligent Document Parsing​

  • Structure Preservation: Maintains document hierarchy, headings, and logical sections
  • Multi-Modal Support: Extracts and processes text, tables, and images
  • RAG-Optimized Output: Delivers chunks with metadata, relationships, and semantic context

Supported File Formats​

FormatExtensionDescription
PDF.pdfAdobe Portable Document Format
Word.docxMicrosoft Word documents
Excel.xlsxMicrosoft Excel spreadsheets
PowerPoint.pptxMicrosoft PowerPoint presentations

Developer Experience​

  • Simple REST API: A small set of endpoints for parsing jobs and retrieval queries
  • Async Processing: Submit jobs and retrieve results when ready
  • Multiple Language Support: Code examples in cURL, Python, and Node.js

How It Works​

  1. Create a Job: Submit a parsing request with your document URL or request an upload URL
  2. Upload (if needed): For local files, upload directly to our secure storage using the presigned URL
  3. Poll or Wait: Check job status or wait for completion
  4. Get Results: Download structured results as a ZIP package containing chunks, images, and tables

Base URL​

All API requests should be made to:

https://api.knowhereto.ai

Need Help?​