htmlbook

htmlbook is an online library designed to host, organize, and share HTML and Markdown documents generated by AI agents or users. It addresses the challenge of moving files such as reports, specifications, and notes off local machines and into a structured, themed, and easily accessible digital bookshelf. Documents are organized by project, grouped with folders and tags, and versioned automatically on every update, providing a searchable and self-sorting library for various document types, including product requirements, study guides, dashboards, recipes, trip plans, and meeting notes.

The platform enables users or AI agents to publish documents directly via Model Context Protocol (MCP), a REST API, or through a command-line interface, with one-time setup for integration. Once documents are pushed to htmlbook, they are re-themed for comfortable reading on any device, including mobile, and can be accessed with a shareable link. Every document remains editable within the reader; users can tap any text to make changes, with each save creating a new version while retaining the original, and a full version history is maintained. The tool supports both standalone HTML documents with custom CSS and JS, as well as Markdown files.

Privacy and data control are central features. By default, all shelves are private, visible only to the owner and explicitly invited team members. Documents are only made public if the user chooses to generate a shareable link, and this can be reversed at any time, removing the public link. Access is authorized server-side for every request, and documents are encrypted both in transit and at rest using TLS, Cloudflare R2, and Google Cloud. API keys are stored as SHA-256 hashes, scoped per workspace, and can be rotated or revoked. Users retain full ownership of their data, with the ability to download originals or permanently delete documents, and the platform commits to never using documents for model training or sharing them without consent.

evidence_sufficient=true