Oxynote

What Oxynote MCP server does

The Oxynote MCP server connects an AI agent to a self-hosted Oxynote workspace for reading, writing, and querying documentation. It is intended for teams that keep operational knowledge, internal API references, commands, and metric explanations together rather than splitting them across a wiki and monitoring system.

Oxynote pages can contain ordinary written content alongside live metric blocks. These blocks support time series, bar charts, and gauges built from one or more Prometheus or SQL queries. SQL sources include PostgreSQL, MySQL, and MariaDB. A metric grid places multiple metric blocks together on a page, allowing the surrounding text to explain thresholds, labels, and operational actions.

The workspace also supports API-style documentation. The README describes split documentation layouts with response fields, request examples, endpoint titles, and live latency charts. Pages can be placed under review, with changes collected into a draft and shown in a block-aware diff before approval.

How it works

Oxynote runs as a self-hosted application. The documented quick start uses a single Docker container, exposes the application on port 8080, and stores data in a mounted volume. After startup, users open the web interface and sign in with the supplied initial administrator credentials.

Within the application, pages can be edited collaboratively. Editors can see other users on a page, identify the block being edited, and add comments to content. Reviewable pages route subsequent changes through a draft instead of allowing direct edits to the protected main version. Reviewers can approve the draft, after which it is merged into main.

For agents, the Oxynote MCP server provides the documented read, write, and query access to those docs. The supplied material does not specify transport details, individual MCP tool names, authentication settings, or a separate client configuration format.

Tools and capabilities

The documented MCP capabilities are:

• Read Oxynote documentation.
• Write or update Oxynote documentation.
• Query Oxynote documentation.

The surrounding Oxynote workspace adds these content features:

• Live metric blocks using Prometheus or SQL queries.
• Time series, bar chart, and gauge visualizations.
• Metric grids containing multiple blocks.
• Split API documentation with request and response content.
• Drafts, block-aware diffs, approvals, and protected main pages.
• Live collaborative editing and comment threads.

Setup and configuration

Start Oxynote with Docker:

docker run -d --name oxynote \
  -p 8080:8080 \
  -v oxynote_data:/oxynote/data \
  ghcr.io/oxynote/oxynote:latest

Open http://localhost:8080 after the container starts. The quick-start instructions provide admin@example.com and oxynote-admin-1234 as the initial login details. Source-based setup is directed to the repository's contribution documentation.

The provided material does not list environment variables or explain how to register the MCP server with a particular desktop or coding client. MCP client configuration should therefore be confirmed from the project's current documentation before deployment.

Limitations and notes

The README describes the documentation workspace and its MCP read, write, and query functions, but does not enumerate individual MCP tool names, MCP transport options, permissions, or authentication behavior. It also does not name supported MCP clients. Those details are important for production integration and are not established by the supplied material.