DOCSEARCHbyAlgolia
Submit Docs

Vectorize API Documentation

website·Indexed 36 days ago

https://docs.vectorize.io

API reference for Vectorize, featuring endpoints for managing source, destination, and AI platform connectors, as well as pipeline orchestration, content extraction, and workspace management.

Tokens
95.2K
Snippets
234
Records
740
Agent score
68%

What's inside Vectorize

  1. Overview of Vectorize RAG pipeline features

    Vectorize provides tools to automate the construction and maintenance of Retrieval-Augmented Generation (RAG) pipelines. Key capabilities include:

    • RAG Pipeline Builder: A user interface to populate vector search indexes from documents, SaaS platforms, and knowledge bases with automatic synchronization to prevent stale data.
    • Advanced Retrieval Endpoint: A built-in endpoint that handles query vectorization, k-ANN search, LLM-based query rewriting based on conversation history, result re-ranking, and providing relevancy scores, cosine similarity, and metadata.
    • Real-Time Updates: Configuration options to immediately update vector search indexes when changes occur in unstructured data sources.
    • Vector Database Integrations: Support for storing embedding vectors in existing databases (e.g., Pinecone, Couchbase) using models from providers like OpenAI and Voyage AI.
    • Management API (Beta): Programmatic control over connectors, API platforms, vector databases, and pipelines for integration into AI applications or CI/CD workflows.

  2. Overview of Vectorize Iris extraction capabilities

    Vectorize Iris is a model-based extraction solution that combines extraction and chunking into a single process. It is designed for complex documents (like PDFs) and provides:

    • Preserved document structure and formatting.
    • Intelligent handling of tables and images.
    • Advanced metadata extraction.
    • Document classification and section identification.

  3. Overview of Vectorize Connectors for Agentic AI

    Vectorize uses connectors to create data pipelines that transform unstructured data into structured, metadata-rich context. This enables AI agents to reason across multiple data sources, use metadata for intelligent filtering (e.g., prioritizing recent documents), perform multi-step analysis, and trigger workflows based on retrieved context.

  4. Overview of Vectorize MCP Server

    The Vectorize MCP server implements the Model Context Protocol (MCP), an open standard that enables AI assistants (such as Claude) to securely access external tools and data sources. It specifically allows AI assistants to perform Vector Retrieval and Deep Research capabilities on private data hosted within Vectorize.

  5. Overview of Vectorize Iris extraction

    Vectorize Iris is a model-based extraction solution that combines document extraction and chunking into a single process. It converts complex documents (PDFs, images, scans, and embedded content) into clean markdown, maintaining text semantics for LLM context. Key capabilities include precise table parsing, support for 100+ languages (including Hindi, Arabic, and Chinese), and structured metadata extraction using JSON schemas.

  6. Overview of White-Labeled OAuth Connectors

    White-Labeled OAuth connectors allow developers to integrate multiple users' cloud storage and productivity platform accounts using their own OAuth applications. This provides full control over branding, the authentication experience, and security compliance while using Vectorize for user management and file processing. This feature is available on the Pro plan and above.

  7. Overview of Google Drive Multi-User (White Label) Connector

    The Google Drive Multi-User (White Label) connector enables applications to integrate multiple users' Google Drive accounts as data sources. It allows developers to maintain brand consistency by using their own Google Cloud OAuth client for authentication and controlling the user experience, while leveraging Vectorize's platform for user management, file processing, and authorization.

  8. Overview of Vectorize agentic data platform

    Vectorize is an agentic data platform that connects AI applications (such as chatbots, AI agents, and automated workflows) to an organization's internal knowledge. It solves common AI data challenges including hallucinations, outdated information, and the inability to process complex documents containing tables, images, and diagrams.

  10. Overview of Dropbox Multi-User (White Label) Connector

    The Dropbox Multi-User (White Label) connector allows applications to integrate multiple users' Dropbox accounts as data sources using the application's own Dropbox App credentials. This ensures brand consistency during authentication, provides control over the user experience, and leverages Vectorize for user management and file processing.

  11. Overview of Vectorize Quickstart Guides

    Vectorize quickstarts are 5-10 minute guides designed to establish a foundation for AI agents to reason over data. Every quickstart follows a three-step process:

    1. Vector Database Setup: Configuring the database with required settings.
    2. Data Pipeline Creation: Building a pipeline to transform content into structured, searchable context.
    3. Data Source Connection: Linking a source (S3, Web Crawler, or Intercom) to the pipeline.

  12. Overview of Vectorize Deep Research

    Deep Research (currently in Beta) generates comprehensive research reports by querying the vector database associated with a specific pipeline. It can augment private vectorized data with information from the web. Note that while there is no additional charge for the feature itself, it utilizes the retrieval endpoint, and these retrievals count against your plan's limits and may trigger additional charges.