What kind of sources does the Deep Research model use?

The model focuses on high-authority sources including academic databases, peer-reviewed journals, meta-analyses, and official government or regulatory reports.

Do I need a Perplexity API key to use this skill?

Yes, you must have a Perplexity API key. You can generate one at perplexity.ai/settings/api and add it to your environment variables as PERPLEXITY_API_KEY.

How long does a research query typically take?

Research typically takes between 30 to 120 seconds, though highly complex queries can take up to 10 minutes. The skill is configured with a 10-minute timeout to accommodate deep analysis.

What happens if the research request times out?

The skill includes an automatic retry mechanism that will attempt the request up to 3 times with exponential backoff. It also auto-saves partial logs to help you troubleshoot.

Can I save the results directly to a specific file?

Yes, you can use the --output flag to specify a file path, or rely on the default auto-save feature which creates timestamped markdown files in your current directory.

Perplexity Deep Research Automation

Name: Perplexity Deep Research Automation
Author: yudame

byyudame

•

データサイエンスとML

Automates deep academic research using the Perplexity Deep Research API to generate cited, evidence-based reports.

This skill integrates Perplexity's sonar-deep-research model directly into your workflow to automate complex, multi-step research tasks. It systematically queries academic databases, official government sources, and peer-reviewed journals to synthesize findings into structured markdown reports complete with inline citations. Built for reliability, the skill manages long-running API requests with a 10-minute timeout, provides automatic retries for transient failures, and ensures data integrity through a timestamped auto-save system for both results and logs.

主な機能

01Timestamped auto-save functionality to prevent data loss during long queries

02Robust error handling with automatic retries and 10-minute timeouts

031 GitHub stars

04Structured markdown output with inline citations and source links

05Automated deep research using the sonar-deep-research model

06Synthesis of peer-reviewed journals and official regulatory sources

ユースケース

01Automating literature reviews and sourcing peer-reviewed data for documentation

02Conducting Phase 1 academic research for podcast episodes and educational scripts

03Generating evidence-based market analysis or technical landscape reports

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yudame/research perplexity-deep-research

For use in Claude.ai and ChatGPT

Download Skill