search_web_pages and fetch_page_content from the terminal or a script, with no API key. mcporter ships a ready-made Keenable config in its repository.
→ mcporter’s Keenable guide · The example config
Install
mcporter needs Node 24 or newer for npm installs:npx in front of each mcporter command below. On macOS, brew install steipete/tap/mcporter also works.
Add the server
Save this asmcporter.json. It registers a server named keenable at https://api.keenable.ai/mcp, and imports: [] keeps mcporter from pulling in servers from your editors’ configs:
mcporter.json
X-Keenable-Title header names the calling app. To use Keenable next to other servers, copy the keenable entry into your existing mcpServers map instead, and keep your own imports setting.
List the tools to check the connection:
search_web_pages and fetch_page_content with their arguments, such as site and published_after.
Search and fetch
Search for sources, then fetch one page as markdown:--output json when a script reads the results. Every argument is in the tool reference.
Add an API key
A key removes the hourly cap. Add it as an environment variable placeholder. mcporter fills it in when it connects, and stops with the variable’s name if it’s unset:mcporter.json
Without a key, calls run on the shared public tier at lower rate limits.
Related pages
NVIDIA NeMo Agent Toolkit
Give NVIDIA NeMo Agent Toolkit agents web search and page fetch with the Keenable function group (keenable__search, keenable__fetch).
Mastra
Give Mastra agents web search and page fetch with the @keenable/mastra package: the keenable-search and keenable-fetch tools.
OctoTools
Give OctoTools solvers web search with Keenable_Search_Tool, a tool card in the OctoTools toolbox, with site and date filters and no API key.
Tool calling
Give an LLM web search and page fetch as tools. Pass the SDK’s TOOLS to any OpenAI-compatible chat completions client and run each call with run_tool_call.