webspec-index 0.4.0

Query WHATWG/W3C web specifications for AI agents and developers

webspec-index

Query WHATWG/W3C web specifications from the command line, Python code, or AI agents (MCP).

Features

• Full-text search across HTML, DOM, URL, and other specifications
• Cross-reference tracking (incoming/outgoing references between specs)
• Fast SQLite-based indexing with FTS5 for instant queries
• Three interfaces: CLI, Python library, and MCP server for AI agents

Installation

pip install webspec-index

Or run directly with uvx (no installation needed):

uvx webspec-index query HTML#navigate

If you install via pip, the webspec-index command is available globally. With uvx, prefix every command with uvx webspec-index instead.

The examples below assume pip install.

Quick Start

Command Line

# Query a specific section
webspec-index query HTML#navigate

# Search across all specs
webspec-index search "tree order" --spec DOM

# Check if a section exists (exit code 0 = found, 1 = not found)
webspec-index exists HTML#navigate

# Find anchors by pattern
webspec-index anchors "*-tree" --spec DOM

# List all headings
webspec-index list HTML

# Get cross-references
webspec-index refs HTML#navigate --direction incoming

# Update to latest spec versions
webspec-index update --spec HTML

# Clear local database
webspec-index clear-db

Most commands support --format json (default) or --format markdown.

Python Library

import webspec_index

# Query a section
result = webspec_index.query("HTML#navigate")
print(result["title"])  # "navigate"
print(result["section_type"])  # "Algorithm"

# Search
results = webspec_index.search("tree order", spec="DOM", limit=5)
for r in results["results"]:
    print(f"{r['spec']}#{r['anchor']}: {r['snippet']}")

# Check existence
if webspec_index.exists("HTML#navigate"):
    print("Section found!")

MCP Server (AI Agents)

Start the MCP server for use with Claude Code or other AI agents:

claude mcp add webspec-index -- uvx webspec-index mcp

Available Specifications

This should support most specs from WHATWG, W3C and TC39.

How It Works

1. Fetches spec HTML from WHATWG/W3C GitHub repositories
2. Parses sections, algorithms, IDL definitions, and cross-references
3. Indexes in SQLite with FTS5 for fast full-text search
4. Tracks versions using git commit SHAs for reproducibility

Development

Built with:

• Rust for fast parsing and indexing (scraper, rusqlite, reqwest)
• PyO3 for zero-cost Python bindings
• Maturin for packaging
• Click for CLI
• MCP (Model Context Protocol) for AI agent integration

License

MIT

Links

• GitHub Repository
• Issue Tracker