Vai al contenuto
AI.info

tools

Unstructured

Unstructured turns complex documents and other files into structured data for search, RAG, analytics, and AI agents.

Unstructured

In inglese

Unstructured processes documents from sources such as S3, SharePoint, Slack, Google Drive, and other connectors. It extracts text, tables, metadata, layout, and images, then supports chunking, enrichment, embeddings, and delivery to storage or vector databases.

Teams use it for document ingestion, RAG pipelines, agent workflows, compliance, and structured data extraction. The free plan includes 10,000 pages to start; Pay-As-You-Go charges $0.015 per page after that, while private deployments use custom pricing.

Features

  • Processes 50+ document and image file types
  • Connects to 40+ sources and 20+ destinations
  • Extracts tables while preserving rows, columns, headers, and relationships
  • Uses OCR for scans, images, handwriting, and low-quality documents
  • Preserves reading order, hierarchy, sections, and page relationships
  • Supports partitioning, chunking, enrichment, and embedding pipelines
  • Provides REST API access for connectors, workflows, and jobs
  • Supports cloud-hosted, dedicated instance, VPC, and bare-metal deployment

Use cases

  • Build RAG pipelines from PDFs, emails, spreadsheets, and presentations
  • Process invoices and compare extracted line items with purchase orders
  • Prepare contracts, filings, and reports for search and analysis
  • Feed structured document data into vector databases and data warehouses
  • Run document-processing tools inside MCP-compatible AI agents

Pros

    Cons

      Latest updates

      • 0.27.0 (0.27.0)

        Enabled telemetry by default, added partition runtime telemetry, and fixed PDF decoding, HTML merging, and charset detection.

      • 0.25.0 (0.25.0)

        partition_json() and partition_ndjson() now handle any valid JSON or NDJSON payload, returning non-Unstructured payloads as Text elements.

      • 0.24.0 (0.24.0)

        The v2 HTML parser now derives category_depth from heading level; URL fetching uses host validation and default timeouts.

      Capabilities

      • Connects to your data — “With 30+ built-in connectors, Unstructured pulls content from your systems of record and business applications—no custom code required.” source
      • Real-time data — “All your data — the messiest, the most difficult, the real-time — transformed with human-level accuracy.” source
      • API — “Use API for full programmatic control.” source
      • Self-hosted — “Unstructured fits your infrastructure—not the other way around. Choose fully managed SaaS, hybrid deployments, VPC installs, or even bare metal.” source

      Get it

      Security

      • SOC 2 Type II — “SOC 2 Type 2” source
      • SOC 2 — “SOC 2 Type 2” source
      • ISO 27001 — “ISO/IEC 27001:2022” source

      Pricing

      Starting price
      Free
      Prices checked
      2026-09-25

      Free

      Free

      • 10,000 free pages to start
      • No card required
      • All features included

      Pay-As-You-Go

      • $0.015 / page
      • 10,000 free pages when your account starts
      • $0.015 per page after the first 10,000 pages
      • All features included

      Business

      Price on request

      • Multi-user accounts
      • All features
      • Dedicated Instance, VPC or Multi-Tenant SaaS
      • Pricing tailored to your needs
      Official website