web-search — Claude Code MCP server: what it does, how to install
Searches the Keenable web index and retrieves page text formatted as clean Markdown. Works without an API key for up to 1,000 requests per hour.
claude mcp add --transport http web-search 'https://api.keenable.ai/mcp?keenable_title=mcp-registry'What web-search does
This server connects your assistant to the Keenable web index for live online queries. It exposes two functions: search_web_pages filters results by site or date, while fetch_page_content extracts clean Markdown up to a default limit of 50000 characters.
Setup does not require hosting search infrastructure. Most clients connect directly to the hosted streamable HTTP endpoint. For editors limited to stdio processes, the @keenable/mcp bridge package routes requests to the cloud service.
Usage is keyless out of the box with a quota of 1,000 requests per hour. If your project exceeds that ceiling, you must supply an optional token via KEENABLE_API_KEY.
What it helps with
- Run live web lookups directly from the agent terminal
- Extract web page text formatted directly into clean Markdown
- Filter online search results by specific domains and publish dates
- Connect web crawling tools without mandatory API token generation
Example prompts for the agent
- Use
search_web_pagesto find recent changes in the Python documentation restricted to python.org - Fetch the technical guide content as clean Markdown using
fetch_page_content - Search for microservice architecture discussions published after January 2024
How to install web-search
1Platform CLI: remote server
claude mcp add --transport http web-search 'https://api.keenable.ai/mcp?keenable_title=mcp-registry'2Platform CLI
claude mcp add --env KEENABLE_API_KEY=<KEENABLE_API_KEY> --transport stdio web-search -- npx -y @keenable/mcpReplace <…> with your own values
1Platform CLI: remote server
codex mcp add web-search --url 'https://api.keenable.ai/mcp?keenable_title=mcp-registry'2Platform CLI
codex mcp add web-search --env KEENABLE_API_KEY=<KEENABLE_API_KEY> -- npx -y @keenable/mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"web-search": {
"url": "https://api.keenable.ai/mcp?keenable_title=mcp-registry"
}
}
}{
"mcpServers": {
"web-search": {
"command": "npx",
"args": [
"-y",
"@keenable/mcp"
],
"env": {
"KEENABLE_API_KEY": "<KEENABLE_API_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"web-search": {
"type": "http",
"url": "https://api.keenable.ai/mcp?keenable_title=mcp-registry"
}
}
}{
"servers": {
"web-search": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@keenable/mcp"
],
"env": {
"KEENABLE_API_KEY": "<KEENABLE_API_KEY>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http web-search 'https://api.keenable.ai/mcp?keenable_title=mcp-registry'2Platform CLI
gemini mcp add -e KEENABLE_API_KEY=<KEENABLE_API_KEY> web-search npx -y @keenable/mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about web-search
How do I install web-search in Claude Code?
Run the command claude mcp add --transport http web-search 'https://api.keenable.ai/mcp?keenable_title=mcp-registry'.
What does the web-search server do?
It gives your assistant web search capabilities over the Keenable index and downloads pages directly as clean Markdown.
Is an API key required to use web-search?
No, the server operates without a key by default and allows 1,000 requests per hour. A key is only needed to lift that limit.
