parallel-cli

Run asynchronous web research jobs with JSON output via the Parallel CLI.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/MarbleSodas/Mavis --skill parallel-cli-marblesodas
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: parallel-cli
Source: https://github.com/MarbleSodas/Mavis/tree/main/optional-skills/research/parallel-cli
Command: npx skills add https://github.com/MarbleSodas/Mavis --skill parallel-cli-marblesodas

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

parallel-cli helps you gather and structure web research results at scale, including extraction, deep multi-step research, enrichment, entity discovery, and ongoing monitoring, without manually coordinating multiple tools.

Core Features & Use Cases

  • Web search with structured output: Perform current lookups and return machine-readable JSON for reliable downstream summarization.
  • Targeted extraction from URLs: Pull clean content or specific slices from web pages using objective-driven extraction.
  • Deep research and enrichment workflows: Run longer research jobs asynchronously, then transform findings into enriched datasets (including CSV/JSON-style column inference).
  • Entity discovery and monitoring: Find large sets of entities (FindAll) and track changes over time (Monitor).

Use Case: You need a set of AI coding-agent startups that offer enterprise capabilities, then enrich the dataset with CEO and annual revenue using non-interactive JSON workflows.

Quick Start

Use parallel-cli search to retrieve structured results in JSON for your exact research question and constraints.

Frequently Asked Questions about parallel-cli

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run async web research jobs and get JSON output?

Async web research jobs launch non-interactively via the CLI and return JSON-first output. You can start jobs, poll for status updates, and chain context across multi-step research runs.

What is entity discovery and how does it work for web extraction?

Entity discovery uses a FindAll mechanism to locate large sets of entities from web sources. It structures the extracted findings into machine-readable datasets for immediate downstream use.

Can I enrich a dataset with CEO and revenue data using CLI web search?

Yes, you can perform data enrichment by running targeted web search workflows that append specific fields like CEO names or annual revenue to your existing tabular datasets.

How do I monitor web pages for changes over time without manual checks?

Recurring change detection monitoring tracks specific web sources over time asynchronously. The CLI polls for updates and returns structured JSON alerts when changes are detected.

Does parallel-cli work for one-shot web lookups or only long research runs?

The tool handles both one-shot lookups and long asynchronous research runs. One-shot queries return structured JSON immediately, while complex workflows launch background jobs for polling.

What's the best way to extract specific content slices from URLs at scale?

Objective-driven extraction pulls clean content or specific slices from web pages at scale. It operates non-interactively and returns structured JSON with disciplined citation URLs.