SiYuan Note is a powerful knowledge management tool used by thousands of professionals to organize their projects, tasks, and data in the form of relational databases (Attribute Views).

The MCP (Model Context Protocol) from Anthropic allows AI assistants like Claude, Cursor, etc. to interact directly with third-party applications. The goal: to make SiYuan fully controllable by AI.

Existing MCP servers for SiYuan manage documents and text blocks, but none support Attribute Views, SiYuan's relational database system.

Specifically:
it is impossible to ask Claude to read, create, or update an entry in a SiYuan database. The most structured part of the tool remained invisible to the AI.

Technical reason:
the official HTTP API of SiYuan is incomplete for write operations on Attribute Views. Data is stored in independent JSON files, not exposed by standard endpoints.

Development of a complete MCP server in TypeScript, published as open source.

Technical approach:
bypassing the incomplete API via the endpoints `/api/file/getFile` and `/api/file/putFile`, direct access to the JSON files of the databases, without access to the local file system. 100% HTTP, compatible with remote instances and Docker.

- Full read of databases (typed columns + all rows)

- Creation of rows with initial values

- Updating multiple cells in a single call

- Filtering rows by column/value

- Programmatic database creation

- Support for 11 column types in read/write

- Auto-discovery of the SiYuan port

```bash

npx siyuan-query-mcp@latest

```

No local compilation required. The user configures Claude Desktop in 3 lines.

- Public npm package: `siyuan-query-mcp`

- Published and maintained on [npmjs.com]
(https://www.npmjs.com/package/siyuan-query-mcp)

- Automated CI/CD pipeline: 1 git tag → GitHub Release + npm publish in 16 seconds

- SiYuan community notified

- Unique feature: only MCP server to support Attribute Views in read/write

GitHub : https://github.com/MyrkoF/siyuan-mcp
npm : https://www.npmjs.com/package/siyuan-query-mcp

- Ability to identify a gap in an open source ecosystem and fill it independently

- Rigorous engineering approach: analysis of the official API, identification of limitations, documented technical bypass

- Mastery of the complete cycle: development → packaging → publication → CI/CD automation → community communication

- Collaboration between humans and AI: development assisted by Claude Code, supervision and technical decisions made by the developer

Tags: projects, MCP, SiYuan, TypeScript, open source, AI, automation, npm, GitHub Actions