context_search and context_scrape tools. An optional OAuth MCP connection adds crawling, extraction, document parsing, brand intelligence, screenshots, monitors, and batches.
Install Context.dev from ClawHub
Review the current package, compatibility, and source before installing.
Choose an access level
You can use either path or enable both.
Install the plugin
1
Install from ClawHub
2
Configure the API key
Create a secret key in the Context.dev dashboard, then set it as If OpenClaw runs as a background service, container, or process manager, set the variable in that service’s environment and restart it there. A shell-only export will not update an already-running Gateway.
CONTEXT_API_KEY in the environment that starts the OpenClaw Gateway.For a Gateway started from the same shell:3
Test native search and scraping
Send the agent a read-only request:For a known page, ask it to use
context_scrape instead.Connect the full MCP catalog
The plugin’s native tools cover search and single-page scraping. Add the hosted MCP server when the agent also needs extraction, parsing, brand data, monitors, screenshots, or batches.Use the integration
Troubleshooting
Reference
Plugin source
README, package metadata, manifest, and release source.
OpenClaw documentation
Gateway, plugins, environment, and MCP configuration.
Context.dev MCP guide
Hosted MCP capabilities and OAuth behavior.
Context.dev dashboard
Create an API key and review usage.