# wso2-docs-mcp-server [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/iamvirul/wso2-docs-mcp-server  
**GitHub Stars:** 3  
**npm Downloads (last month):** 596  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/wso2-docs-mcp-server

## Description
Semantic search over WSO2 docs (APIM, MI, Choreo, Ballerina) via RAG and pgvector.

## Tools
Capabilities this server exposes over MCP:

- **search_wso2_docs** — Semantic search across all products. Optional `product` and `limit` filters.
- **get_wso2_guide** — Search within a specific product (`apim`, `mi`, `bi`, `choreo`, `is`, `ballerina`, `library`).
- **explain_wso2_concept** — Broad concept search across all products, returns 8 top results.
- **list_wso2_products** — Returns all supported products with IDs and base URLs.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "wso2-docs-mcp-server": {
    "command": "npx",
    "args": ["-y","wso2-docs-mcp-server"],
    "env": {
      "DATABASE_URL": "",
      "EMBEDDING_PROVIDER": ""
    }
  }
}
```

**Requires environment variables:** `DATABASE_URL`, `EMBEDDING_PROVIDER` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What wso2-docs-mcp-server MCP server does

The wso2-docs-mcp-server MCP server gives MCP-compatible AI clients semantic access to indexed WSO2 documentation. Its source catalog covers API Manager, Micro Integrator, Ballerina Integrator, Choreo, Identity Server, Ballerina, and the WSO2 Library. Search results can include a title, excerpt, source URL, product identifier, section, and relevance score.

The project is an unofficial community implementation and is not affiliated with or endorsed by WSO2. It is useful for assistants helping with WSO2 configuration, deployment, product concepts, and documentation lookup across multiple product families.

## How it works

Documentation is collected by a dual ingestion process. For products with suitable public documentation repositories, the crawler reads raw Markdown through GitHub's Git Trees API. Products without dedicated GitHub documentation repositories can be handled through the web-crawl fallback. The resulting content is embedded and stored in PostgreSQL with the pgvector extension.

Embeddings use Ollama by default, with the `nomic-embed-text` model shown in the setup instructions. If Ollama is unavailable, the server can use HuggingFace ONNX inference in process and download the model when needed. OpenAI, Google Gemini, and Voyage AI are also supported as cloud embedding providers.

The MCP process runs on demand through an AI client's standard input/output transport. A PostgreSQL connection must point the server and its indexing commands at the same database. Database migration should be run before crawling, and again if the embedding dimensions change after switching providers.

## Setup and configuration

The wso2-docs-mcp-server MCP server requires Node.js 20 or newer and Docker for the pgvector database. Install it globally from npm, or clone the repository and build it locally. Start the supplied PostgreSQL/pgvector Compose service, run the migration command, and index either all products or a selected product. The crawler also supports a result limit for focused tests and a force option for re-indexing unchanged pages.

The primary connection setting is `DATABASE_URL`. `EMBEDDING_PROVIDER` selects the embedding backend; the documented default is Ollama. Cloud providers require their corresponding API key, while the local Ollama and HuggingFace paths do not require an API key. After indexing, configure the MCP command in Claude Desktop, Claude Code, Cursor, or VS Code. An npm installation can invoke `wso2-docs-mcp-server`; a source checkout can launch the built Node entry point.

## Tools and capabilities

The wso2-docs-mcp-server MCP server exposes four tools:

- `search_wso2_docs` searches across the complete documentation collection and accepts optional product and result-limit filters.
- `get_wso2_guide` restricts a search to one supported product ID, such as `apim`, `mi`, `choreo`, or `ballerina`.
- `explain_wso2_concept` performs a broad cross-product concept search and returns eight top results.
- `list_wso2_products` returns the supported product IDs and their base URLs.

Use the product-specific tool when answers must stay within one WSO2 product. Use the cross-product tools when a concept may span API management, integration, identity, or Ballerina documentation.

## Limitations and notes

The repository describes this as an unofficial project, so it should not be treated as an official WSO2 support channel. Search quality depends on the local index and embedding configuration; documentation must be crawled before it can be retrieved. The server also depends on a running PostgreSQL/pgvector service and a completed migration. Changing embedding providers may require migration because vector dimensions can differ. The documented product list and crawl sources should be checked when current documentation coverage matters.

_Full upstream README: https://allmcps.com/mcp/wso2-docs-mcp-server/readme_