PageIndex JS SDK
https://github.com/vectifyai/pageindex-js-sdk- Category
- AI Tools
- Rank
- No. 1899Tools index
Previous survey · No. 1870 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- VectifyAI
- GitHub
- 7 stars
- Latest release
- v0.8.0
- Date
About
TypeScript SDK for PageIndex document processing.
What it does
PageIndex JS SDK connects Node applications to a hosted document service. It uploads files, exposes document metadata and tree structures, manages folders, and supports conversational queries across one or several PDFs. Developers can use direct REST operations or typed MCP tool wrappers for agent workflows.
Why it's ranked here
Its value comes from covering both conventional application code and agent-oriented document access behind one client. Streaming chat, citations, folder scoping, structured errors, page retrieval, image retrieval, and document discovery form a coherent toolkit. The main constraint is dependence on the hosted PageIndex API and an API key.
What's good
The SDK provides typed results for documents, folders, trees, chat responses, streaming chunks, and MCP operations. Chat supports multiple documents, citations, and metadata during streaming. MCP connections open lazily, close after a configurable idle period, and reconnect when folder scope changes. Errors preserve codes, details, and HTTP status where available.
Tradeoffs
This is a client for a remote service, not a local PDF processor. The REST surface and MCP surface overlap but expose different capabilities and naming models. Stream parsing expects server-sent event lines with a specific data prefix and valid JSON payloads. Several older MCP operations remain callable only for backward compatibility and may disappear later.
How to use it well
Choose it for a Node application that already wants PageIndex-hosted ingestion, retrieval, and PDF chat. Use REST operations for uploads, metadata, folders, trees, and chat. Use the MCP wrappers when an agent needs document discovery, page text, structure, or images. Bring another component for local parsing, storage, or model execution.
Technical notes+
package.json defines an ESM package with separate import and require exports, declaration outputs for both consumers, Node 18 or newer, and one runtime dependency on the Model Context Protocol SDK. src/client.ts composes PageIndexApi with a lazily created PageIndexTools instance and supports asynchronous disposal. src/api/client.ts uses fetch, FormData, typed overloads for streamed versus non-streamed chat, and a line-buffered SSE parser. src/transport.ts uses StreamableHTTPClientTransport, bearer authentication, optional X-Folder-Scope headers, automatic connection, and idle closure. src/errors.ts defines the shared error taxonomy. src/tools/index.ts retains three deprecated wrappers for backward compatibility.
Observed
- License
- MIT
- Primary language
- TypeScript
- Installation
- Published as @pageindex/sdk and installed through npm
- Interfaces
- Node library exposing REST API operations and typed MCP tool wrappers
- Runtime support
- Node.js 18 or newer
- Package formats
- Separate ESM import and CommonJS require exports with TypeScript declarations
- Runtime dependencies
- One declared dependency: @modelcontextprotocol/sdk
Read from README.md, package.json, src/index.ts, src/client.ts, src/errors.ts, src/transport.ts, src/api/types.ts, src/api/client.ts, src/tools/index.ts, src/tools/types.ts, src/tools/get-document.ts, src/tools/list-folders.ts, src/tools/remove-document.ts, src/tools/browse-documents.ts, src/tools/get-page-content.ts.
What it can do
Process documents using TypeScript
Documents → Processed document data
Index document content
Document files → Searchable document index
Extract text from documents
Document files → Extracted text content
Parse document metadata
Document files → Document metadata objects
Search indexed documents
Search queries → Matching document results
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.