Agent Skill

Serply publishes a SKILL.md file that teaches a coding agent how to use Serply: which endpoint to call for which job, the X-Api-Key auth header, the MCP tool names, and gotchas like Google Maps taking its parameters differently from every other endpoint.

Once installed, an agent that already speaks the Agent Skills format (Claude Code, claude.ai, or the Claude API) picks it up automatically whenever a task needs live search results or page scraping - no need to paste API docs into the conversation.

Install in Claude Code

npx skills add serply-inc/skills

Or download the file directly into your project's or personal skills directory:

mkdir -p .claude/skills/searching-with-serply
curl -o .claude/skills/searching-with-serply/SKILL.md https://serply.io/skill.md

Use ~/.claude/skills/searching-with-serply/SKILL.md instead of the project-local path to make it available across every project. Claude Code discovers and loads it automatically.

Install in claude.ai

Download the file above, then upload it as a custom Skill under Settings > Features (requires code execution to be enabled on your plan).

What's in it

  • The REST endpoint table (Search, News, Maps, Images, Jobs, Scholar, Product, Video, Bing, Page Fetch) with paths and response shapes
  • The X-Api-Key auth header and the free-credit signup link
  • The hosted MCP server address and its eight tool names
  • The one real gotcha worth knowing before you build against this API: Google Maps takes query parameters after a ? instead of packed into the path, and does not support X-Proxy-Location or X-User-Agent

The full endpoint reference - every parameter, every response field - lives in the API Endpoints section; the Skill links out to it rather than duplicating it, so it stays short enough for an agent to read in one pass.

Source

The Skill's source is github.com/serply-inc/skills, also served at serply.io/skill.md and mirrored in this project at .claude/skills/searching-with-serply/SKILL.md.

View as Markdown