LoginSign Up
← Back to all sparks
D

Document360

COLLAB
Velocity5.0

Document360 builds for AI agents as readers, adding MCP tools, llms.txt, and LLM-structured guides

knowledge-baseai-agentsmcp-serverllms-txtdocs-platformeddy-ai
◆Current state
Document360 is a knowledge base platform shipping consistent monthly releases across its AI (Eddy AI), editor, analytics, and developer API layers. It's building a genuine AI-agent consumption story: automatic llms.txt generation tells AI assistants what the docs cover, MCP server tools let agents query and list articles programmatically, and step-by-step guides are now structured for LLM readability. The recent API v3 rebuild (OAuth 2.0, scoped keys) completes the programmable surface. Collaborative editing, article archiving, and Eddy AI analytics round out the human-facing improvements.
◆Where it's heading
The directional bet here is that documentation will be consumed by AI agents as much as by human readers, and Document360 wants to be the authoritative source agents retrieve from. Every recent release either deepens the Eddy AI layer (analytics on unanswered queries, ticket escalation, reader group controls) or expands the agent-accessible surface (MCP server, llms.txt, LLM-structured content). The API v3 with OAuth 2.0 means external systems can now read and write the knowledge base programmatically. The platform is quietly becoming a knowledge-as-a-service layer.
◆Prediction
The next move is tighter integration between the MCP server and the Eddy AI chatbot — having the AI chatbot call back into the knowledge base via the same programmatic tools that external agents use, closing the loop on self-referential AI reasoning over docs. First-party support for major agent frameworks is likely to follow.

◆Recent moves

  1. 8h ago

    September 12.9.3: Drive renaming, Eddy AI email escalations, and content access clarity

    Document360 12.9.3 ships Drive file renaming without breaking existing references, email routing for Eddy AI ticket escalations, Drive asset insertion in AI capture, and clearer content access inheritance indicators. Solid polish across file management and AI chatbot escalation flows, continuing the steady Eddy AI build-out.

    View source ↗
  2. 7d ago

    September 12.9.2: Article archiving, private link expiry, and new MCP server tools

    12.9.2 adds article archiving (clean removal from search/AI without deletion), private link expiry controls, and two new MCP server tools — full-text search and article listing with regex support. The MCP expansion is the most meaningful piece: it widens the programmatic surface that AI agents can use to query documentation.

    View source ↗
  3. 21d ago

    September 12.9.1: Eddy AI analytics expose unanswered query patterns

    12.9.1 introduces Advanced Eddy AI Analytics — tracking answered vs. unanswered queries with reasons for the gaps (content gap, out-of-scope, query interpretation failure). This gives knowledge base managers data to close documentation holes that the AI is already hitting.

    View source ↗
  4. 28d ago

    August 2026 - 12.8.5

    12.8.5 redesigned the top navigation to consolidate project, workspace, and language switching into one unified dropdown — a meaningful friction reduction for teams managing multilingual or multi-project documentation. Label management tooling (rename, delete, group in bulk) closes a long-standing gap in content organization.

    View source ↗
  5. 1mo ago

    August 2026 - 12.8.4

    12.8.4 ships collaborative editing in the Advanced WYSIWYG editor — multiple contributors see each other's cursors, edits, and comments in real time, with publishing that creates a new draft automatically so work continues uninterrupted. Also notable: step-by-step guides are now structured for AI agent/LLM readability, a quiet but directional update.

    View source ↗
  6. 1mo ago

    August 2026 - 12.8.3

    12.8.3 revamps the Publish dialog with AI-powered metadata suggestions, auto-fix, and pre-publish checks (broken links, unresolved comments) all in one workspace. Offline documentation exports now mirror the live knowledge base layout more faithfully, including table of contents, breadcrumbs, and navigation.

    View source ↗