turath
MCP server to search and retrieve passages from a corpus of 7,872 classical Islamic books via the Sahifah API, with full citations and mu'tabar filtering.
README
turath
Cited ibaroh — raw, sourced passages — from a corpus of 7,872 classical Islamic books (kitab turats: fiqh, usul, hadith commentary, tafsir, and more), pulled through the Sahifah API.
You query in Indonesian, English, or Arabic; the corpus is searched in Arabic; you get back the original passage — not a paraphrase — with its full citation: book, author, table-of-contents path, juz, and page. A mu'tabar filter keeps results to works recognized as authoritative in the tradition.
This repo ships three ways to use it, all built on the same read-only,
revocable Sahifah researcher API key (prefixed bm_):
- A Claude Code plugin bundling the
/turathskill and a remote MCP connector. - A
/turathskill you can drop into any Claude Code project on its own. - A remote MCP server (
turath_search,get_ibaroh) for any MCP-speaking agent — plus the raw HTTP API underneath it.
Adab of naql (the ethic of transmission). Everything here surfaces what the Sahifah corpus indexed; it is not a substitute for opening the actual kitab. Every result ends with a reminder to verify the quotation against the primary source before relying on it — especially for anything bound for a fatwa, an argument, or published writing. Where a passage looks truncated or OCR-garbled, treat it as uncertain. Where the fuqaha differ, note the khilaf rather than picking a side. Allahu a'lam.
Getting an API key
Sign in at sahifah.online and create a key from
the /api-keys dashboard. Keys are read-only, rate-limited, capped per
user, and revocable at any time. The key is never stored or logged by anything
in this repo — it is read from your environment or passed per-request.
export SAHIFAH_API_KEY=bm_your_key_here # bash / zsh
$env:SAHIFAH_API_KEY = "bm_your_key_here" # PowerShell
1. Install as a Claude Code plugin
/plugin marketplace add dihannahdi/turath
/plugin install turath@sahifah
This registers the /turath skill and connects the remote MCP server
(https://sahifah.online/mcp). Set SAHIFAH_API_KEY in your environment first
so the MCP connector can authenticate (see .mcp.json).
2. Use the /turath skill directly
The skill lives in skills/turath/ and drives a stdlib-only Python script —
no dependencies to install.
# search for candidate passages
python skills/turath/scripts/turath.py search "hukum zakat fitrah dengan uang" --max 10
# restrict to mu'tabar (authoritative) kitab only
python skills/turath/scripts/turath.py search "penyakit hasad dalam hati" --mutabar-only
# fetch a full passage by the book_id / toc_id printed above
python skills/turath/scripts/turath.py ibaroh 123 456
See skills/turath/SKILL.md for the full command reference and the adab of
naql the skill carries through its output.
3. Remote MCP server
The server is a stateless Streamable HTTP MCP server exposing two tools:
turath_search— search the corpus; returns each passage's citation, book/author, TOC path, page/juz, madhhab, era, and relevance score.get_ibaroh— given abook_id/toc_idfrom a search, fetch the full passage text and complete citation.
Add it as a Claude connector (or with claude mcp add):
claude mcp add turath --transport http https://sahifah.online/mcp \
--header "X-Sahifah-Api-Key: bm_your_key_here"
The server never hardcodes or stores a key. For each request it resolves the
key to forward from, in order: an Authorization: Bearer <key> header, an
X-Sahifah-Api-Key / X-Api-Key header, or its own SAHIFAH_API_KEY env var
(single-tenant fallback). The key is never logged.
Self-hosting the MCP server
The TypeScript source is in src/ and is what runs behind
https://sahifah.online/mcp. To run your own instance:
npm install
npm run build # tsc -> dist/
npm start # node dist/index.js (listens on PORT, default 8795)
Put a reverse proxy in front for TLS and never expose the raw port publicly.
See the comments in src/index.ts and ecosystem.config.cjs (PM2) for the
deployment shape. Configuration lives in .env (see .env.example):
SAHIFAH_API_BASE is fixed at startup and never taken from request input.
Underlying HTTP API
Both the skill and the MCP server sit on the public Sahifah API:
POST /api/v1/search—{ query, max_results?, mutabar_only?, include_ai? }, auth viaX-API-Key.include_aidefaults tofalse— raw ibaroh first.POST /api/v1/passage—{ book_id, toc_id }→ full passage + citation.
Layout
turath/
├── .claude-plugin/
│ ├── plugin.json # plugin manifest
│ └── marketplace.json # single-plugin marketplace (name: sahifah)
├── .mcp.json # remote MCP connector (auto-discovered by the plugin)
├── skills/turath/
│ ├── SKILL.md # the /turath skill
│ └── scripts/turath.py # stdlib-only client
├── src/ # TypeScript MCP server (Express + @modelcontextprotocol/sdk)
├── package.json
├── tsconfig.json
├── ecosystem.config.cjs # PM2
└── .env.example
Validate before publishing
claude plugin validate .
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.