MCP server
Web scraping for OpenAI Codex
Give Codex reliable web access through MCP.
Codex supports MCP servers in the CLI and IDE extension using shared config.toml configuration. Add webclaw over MCP and Codex can fetch live documentation, scrape protected pages, summarize sources, and extract structured data while it edits your codebase.
You are set up. Now try it.
Check it connected
webclaw should appear in the list of configured servers.
Then paste this
Use webclaw to read the current documentation for the library I am importing in this file, then update my code to match the API it describes rather than the version you remember.
Why webclaw for Codex
- Fits Codex's MCP config.toml format
- Works from both Codex CLI and IDE extension configuration
- Fast scrape and search tools without browser setup
- LLM-ready markdown keeps Codex context focused
Common use cases
- Let Codex inspect current docs before changing code
- Research libraries, API references, and migration guides
- Scrape product pages or support docs into implementation notes
- Build repeatable web-data workflows as coding tasks
Add a [mcp_servers.webclaw] table to ~/.codex/config.toml, or to a project-scoped .codex/config.toml in a trusted project. Codex CLI and the IDE extension share this configuration.