User Guide

Welcome to the Elsai Core User Guide v2.0.0. This section provides comprehensive documentation for the updated Elsai components and modules.

Components Overview

Elsai Models Hub and MCP

• elsai-model-hub: OpenAI-compatible API for hosted models (for example Gemma-4 E4B and Phi-4-mini) with chat, multimodal, and tool-calling examples.

• elsai-mcp: Automatically converts any backend Swagger (OpenAPI) API into a Model Context Protocol (MCP) server, allowing LLMs to interact with APIs as tools.

Document Processing

• elsai-ocr-extractors: In version 2.0.0, VisionAI Extractor has been enhanced with unified file extraction supporting both PDF files and image formats (.png, .jpg, .jpeg, .gif, .bmp, .tiff, .webp). The methods extract_text_from_pdf and extract_text_from_pdf_async have been renamed to extract_text_from_file and extract_text_from_file_async respectively, with the parameter pdf_path changed to file_path to reflect multi-format support.

Cloud and Email Services

• elsai-cloud-connectors: In version 2.0.0, the package introduces EmailService, a comprehensive Gmail SMTP/IMAP connector that supports sending, listing, retrieving, replying, and forwarding emails with both password and OAuth2 authentication.

Getting Started

To get started with Elsai Core v2.0.0, we recommend beginning with elsai-model to understand the enhanced LLM capabilities, which form the foundation for most other components.