How does deep research differ from standard web search?

Standard web search is for quick lookups, while deep research is an exhaustive process that can take 10-100x longer to synthesize dozens of sources into a comprehensive report.

Does this skill require an API key?

Yes, it requires the parallel-cli to be installed and authenticated via a 'parallel-cli login' or by setting the PARALLEL_API_KEY environment variable.

What happens if a research task takes longer than 9 minutes?

The skill includes a polling mechanism that can be re-run if it times out, ensuring that deep research continues server-side until completion even during long-running tasks.

Where are the research results saved?

Results are saved locally to your workspace as two files: a formatted .md report for reading and a .json file containing the metadata and research basis.

What are the different processor tiers available?

The skill offers Pro-Fast (30s-5m), Ultra-Fast (1-10m), and Ultra (5-25m) tiers, allowing you to choose the level of depth and analysis required for your topic.

Parallel Deep Research

Name: Parallel Deep Research
Author: parallel-web

byparallel-web

•

Web Scraping & Data Collection

Conducts exhaustive, multi-step web investigations and generates comprehensive reports for complex research topics.

The Parallel Deep Research skill empowers Claude to move beyond simple lookups and perform high-intensity, multi-minute investigations. By utilizing the parallel-cli, it synthesizes information from dozens of sources to produce structured Markdown reports and metadata-rich JSON files. This skill is specifically designed for scenarios where the user requires a 'thorough investigation' or 'exhaustive report,' offering three tiers of processing depth (Pro-Fast, Ultra-Fast, and Ultra) to balance speed and comprehensiveness. It handles long-running tasks asynchronously, allowing users to continue working while the research process tracks and polls for final completion.

Key Features

0113 GitHub stars

02Built-in polling logic to manage long-running investigative tasks

03Automatic generation of detailed Markdown reports and JSON metadata

04Asynchronous execution with real-time monitoring URLs

05Tiered processing (Pro-Fast to Ultra) for customizable research depth

06Executive summary extraction for rapid insight delivery

Use Cases

01Generating exhaustive market analysis and industry trend reports

02Performing deep-dive technical comparisons between competing software frameworks

03Conducting comprehensive investigative research for professional or academic documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add parallel-web/parallel-agent-skills parallel-deep-research

For use in Claude.ai and ChatGPT

Download Skill