parallel-ai-search

name	parallel-ai-search
description	Web search + URL extraction via Parallel Search/Extract APIs. Use for up-to-date research, domain-scoped searching, and extracting LLM-ready excerpts/markdown from URLs.
homepage	https://docs.parallel.ai/search/search-quickstart
metadata	{"openclaw":{"emoji":"🔎","homepage":"https://docs.parallel.ai/","requires":{"bins":["node"],"env":["PARALLEL_API_KEY"]},"primaryEnv":"PARALLEL_API_KEY"}}

Parallel AI Search (OpenClaw skill)

Use this skill to run web research through Parallel Search (ranked, LLM-optimised excerpts) and Parallel Extract (clean markdown from specific URLs, including JS-heavy pages and PDFs).

The skill ships tiny Node .mjs helpers so the agent can call the APIs deterministically via the OpenClaw exec tool.

Quick start

1) Provide the API key

Prefer configuring it in ~/.openclaw/openclaw.json (host runs):

{
  skills: {
    entries: {
      "parallel-ai-search": {
        enabled: true,
        apiKey: "YOUR_PARALLEL_API_KEY"
      }
    }
  }
}

Notes:

• If the agent run is sandboxed, the Docker sandbox does not inherit host env. Provide the key via agents.defaults.sandbox.docker.env (or bake it into the image).
• This skill is gated on PARALLEL_API_KEY. If it’s missing, OpenClaw won’t load the skill.